Search Results

Search found 17 results on 1 pages for 'fasta'.

Page 1/1 | 1 

  • Converting FASTQ to FASTA with SED/AWK

    - by neversaint
    I have a data in that always comes in block of four in the following format (called FASTQ): @SRR018006.2016 GA2:6:1:20:650 length=36 NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNGN +SRR018006.2016 GA2:6:1:20:650 length=36 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!+! @SRR018006.19405469 GA2:6:100:1793:611 length=36 ACCCGCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC +SRR018006.19405469 GA2:6:100:1793:611 length=36 7);;).;);;/;*.2>/@@7;@77<..;)58)5/>/ Is there a simple sed/awk/bash way to convert them into this format (called FASTA): >SRR018006.2016 GA2:6:1:20:650 length=36 NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNGN >SRR018006.19405469 GA2:6:100:1793:611 length=36 ACCCGCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC In principle we want to extract the first two lines in each block-of-4 and replace @ with >.

    Read the article

  • Java looping through array - Optimization

    - by oudouz
    I've got some Java code that runs quite the expected way, but it's taking some amount of time -some seconds- even if the job is just looping through an array. The input file is a Fasta file as shown in the image below. The file I'm using is 2.9Mo, and there are some other Fasta file that can take up to 20Mo. And in the code im trying to loop through it by bunches of threes, e.g: AGC TTT TCA ... etc The code has no functional sens for now but what I want is to append each Amino Acid to it's equivalent bunch of Bases. Example : AGC - Ser / CUG Leu / ... etc So what's wrong with the code ? and Is there any way to do it better ? Any optimization ? Looping through the whole String is taking some time, maybe just seconds, but need to find a better way to do it. import java.io.BufferedReader; import java.io.File; import java.io.FileNotFoundException; import java.io.FileReader; import java.io.IOException; public class fasta { public static void main(String[] args) throws IOException { File fastaFile; FileReader fastaReader; BufferedReader fastaBuffer = null; StringBuilder fastaString = new StringBuilder(); try { fastaFile = new File("res/NC_017108.fna"); fastaReader = new FileReader(fastaFile); fastaBuffer = new BufferedReader(fastaReader); String fastaDescription = fastaBuffer.readLine(); String line = fastaBuffer.readLine(); while (line != null) { fastaString.append(line); line = fastaBuffer.readLine(); } System.out.println(fastaDescription); System.out.println(); String currentFastaAcid; for (int i = 0; i < fastaString.length(); i+=3) { currentFastaAcid = fastaString.toString().substring(i, i + 3); System.out.println(currentFastaAcid); } } catch (NullPointerException e) { System.out.println(e.getMessage()); } catch (FileNotFoundException e) { System.out.println(e.getMessage()); } catch (IOException e) { System.out.println(e.getMessage()); } finally { fastaBuffer.close(); } } }

    Read the article

  • Efficient file buffering & scanning methods for large files in python

    - by eblume
    The description of the problem I am having is a bit complicated, and I will err on the side of providing more complete information. For the impatient, here is the briefest way I can summarize it: What is the fastest (least execution time) way to split a text file in to ALL (overlapping) substrings of size N (bound N, eg 36) while throwing out newline characters. I am writing a module which parses files in the FASTA ascii-based genome format. These files comprise what is known as the 'hg18' human reference genome, which you can download from the UCSC genome browser (go slugs!) if you like. As you will notice, the genome files are composed of chr[1..22].fa and chr[XY].fa, as well as a set of other small files which are not used in this module. Several modules already exist for parsing FASTA files, such as BioPython's SeqIO. (Sorry, I'd post a link, but I don't have the points to do so yet.) Unfortunately, every module I've been able to find doesn't do the specific operation I am trying to do. My module needs to split the genome data ('CAGTACGTCAGACTATACGGAGCTA' could be a line, for instance) in to every single overlapping N-length substring. Let me give an example using a very small file (the actual chromosome files are between 355 and 20 million characters long) and N=8 import cStringIO example_file = cStringIO.StringIO("""\ header CAGTcag TFgcACF """) for read in parse(example_file): ... print read ... CAGTCAGTF AGTCAGTFG GTCAGTFGC TCAGTFGCA CAGTFGCAC AGTFGCACF The function that I found had the absolute best performance from the methods I could think of is this: def parse(file): size = 8 # of course in my code this is a function argument file.readline() # skip past the header buffer = '' for line in file: buffer += line.rstrip().upper() while len(buffer) = size: yield buffer[:size] buffer = buffer[1:] This works, but unfortunately it still takes about 1.5 hours (see note below) to parse the human genome this way. Perhaps this is the very best I am going to see with this method (a complete code refactor might be in order, but I'd like to avoid it as this approach has some very specific advantages in other areas of the code), but I thought I would turn this over to the community. Thanks! Note, this time includes a lot of extra calculation, such as computing the opposing strand read and doing hashtable lookups on a hash of approximately 5G in size. Post-answer conclusion: It turns out that using fileobj.read() and then manipulating the resulting string (string.replace(), etc.) took relatively little time and memory compared to the remainder of the program, and so I used that approach. Thanks everyone!

    Read the article

  • Can anyone tell me why these lines are not working?

    - by user343934
    I am trying to generate tree with fasta file input and Alignment with MuscleCommandline import sys,os, subprocess from Bio import AlignIO from Bio.Align.Applications import MuscleCommandline cline = MuscleCommandline(input="c:\Python26\opuntia.fasta") child= subprocess.Popen(str(cline), stdout = subprocess.PIPE, stderr=subprocess.PIPE, shell=(sys.platform!="win32")) align=AlignIO.read(child.stdout,"fasta") outfile=open('c:\Python26\opuntia.phy','w') AlignIO.write([align],outfile,'phylip') outfile.close() I always encounter with these problems Traceback (most recent call last): File "", line 244, in run_nodebug File "C:\Python26\muscleIO.py", line 11, in align=AlignIO.read(child.stdout,"fasta") File "C:\Python26\Lib\site-packages\Bio\AlignIO_init_.py", line 423, in read raise ValueError("No records found in handle") ValueError: No records found in handle

    Read the article

  • Subprocess fails to catch the standard output

    - by user343934
    I am trying to generate tree with fasta file input and Alignment with MuscleCommandline import sys,os, subprocess from Bio import AlignIO from Bio.Align.Applications import MuscleCommandline cline = MuscleCommandline(input="c:\Python26\opuntia.fasta") child= subprocess.Popen(str(cline), stdout = subprocess.PIPE, stderr=subprocess.PIPE, shell=(sys.platform!="win32")) align=AlignIO.read(child.stdout,"fasta") outfile=open('c:\Python26\opuntia.phy','w') AlignIO.write([align],outfile,'phylip') outfile.close() I always encounter with these problems Traceback (most recent call last): File "<string>", line 244, in run_nodebug File "C:\Python26\muscleIO.py", line 11, in <module> align=AlignIO.read(child.stdout,"fasta") File "C:\Python26\Lib\site-packages\Bio\AlignIO\__init__.py", line 423, in read raise ValueError("No records found in handle") ValueError: No records found in handle

    Read the article

  • Multiple variables for the same command

    - by Lowzenza
    I'm new to cmd and was wondering if there's an easier way of retpying a variable in a command. For example, I have to do two commands for a set of 96 files and each time I would hit the up arrrow key, get my old commands back and change a variable from 1 to 2, then 2 to 3 and so forth. i.e.: Desktop\InitialProcess_230 Process230input.fasta -output Process230.fasta Then each time I want to do the next file which would be InitialProcess_231 and so on, I would change that in the command by scrolling along and removing 0 and putting a 1. Doing that for almost a 100 files seems like a hassle.

    Read the article

  • File management

    - by user343934
    I am working on python and biopython right now. I have a file upload form and whatever file is uploaded suppose(abc.fasta) then i want to pass same name in execute (abc.fasta) function parameter and display function parameter (abc.aln). Right now i am changing file name manually, but i want to have it automatically. Workflow goes like this. ----If submit is not true then display only header and form part --- if submit is true then call execute() and get file name from form input --- Then display the save file result in the same page. File name is same as input. My raw code is here -- http://pastebin.com/FPUgZSSe Any suggestions, changes and algorithm is appreciated Thanks

    Read the article

  • saving file name from html form input

    - by user343934
    I am working on python and biopython right now. I have a file upload form and whatever file is uploaded suppose(abc.fasta) then i want to pass same name in execute (abc.fasta) function parameter and display function parameter (abc.aln). Workflow goes like this. ----If submit is not true then display only header and form part --- if submit is true then call execute() and get file name from form input --- Then display the save file result in the same page. File name is same as input. My raw code is here -- http://pastebin.com/FPUgZSSe Any suggestions, changes and algorithm is appreciated Thanks

    Read the article

  • On Disk Substring index

    - by emeryc
    I have a file (fasta file to be specific) that I would like to index, so that I can quickly locate any substring within the file and then find the location within the original fasta file. This would be easy to do in many cases, using a Trie or substring array, unfortunately the strings I need to index are 800+ MBs which means that doing them in memory in unacceptable, so I'm looking for a reasonable way to create this index on disk, with minimal memory usage. (edit for clarification) I am only interested in the headers of proteins, so for the largest database I'm interested in, this is about 800 MBs of text. I would like to be able to find an exact substring within O(N) time based on the input string. This must be useable on 32 bit machines as it will be shipped to random people, who are not expected to have 64 bit machines. I want to be able to index against any word break within a line, to the end of the line (though lines can be several MBs long). Hopefully this clarifies what is needed and why the current solutions given are not illuminating. I should also add that this needs to be done from within java, and must be done on client computers on various operating systems, so I can't use any OS Specific solution, and it must be a programatic solution.

    Read the article

  • Deleting While Iterating in Ruby?

    - by Jesse J
    I'm iterating over a very large set of strings, which iterates over a smaller set of strings. Due to the size, this method takes a while to do, so to speed it up, I'm trying to delete one of the strings from the smaller set that no longer needs to be used. Below is my current code: Ms::Fasta.foreach(@database) do |entry| all.each do |set| if entry.header[1..40].include? set[1] + "|" startVal = entry.sequence.scan_i(set[0])[0] if startVal != nil @locations << [set[0], set[1], startVal, startVal + set[1].length] all.delete(set) end end end end The problem I face is that the easy way, array.delete(string), effectively adds a break statement to the inner loop, which messes up the results. The only way I know how to fix this is to do this: Ms::Fasta.foreach(@database) do |entry| i = 0 while i < all.length set = all[i] if entry.header[1..40].include? set[1] + "|" startVal = entry.sequence.scan_i(set[0])[0] if startVal != nil @locations << [set[0], set[1], startVal, startVal + set[1].length] all.delete_at(i) i -= 1 end end i += 1 end end This feels kind of sloppy to me. Is there a better way to do this? Thanks.

    Read the article

  • Parse a text file into multiple text file

    - by Vijay Kumar Singh
    I want to get multiple file by parsing a input file Through Java. The Input file contains many fasta format of thousands of protein sequence and I want to generate raw format(i.e., without any comma semicolon and without any extra symbol like "", "[", "]" etc) of each protein sequence. A fasta sequence starts form "" symbol followed by description of protein and then sequence of protein. For example ? lcl|NC_000001.10_cdsid_XP_003403591.1 [gene=LOC100652771] [protein=hypothetical protein LOC100652771] [protein_id=XP_003403591.1] [location=join(12190..12227,12595..12721,13403..13639)] MSESINFSHNLGQLLSPPRCVVMPGMPFPSIRSPELQKTTADLDHTLVSVPSVAESLHHPEITFLTAFCL PSFTRSRPLPDRQLHHCLALCPSFALPAGDGVCHGPGLQGSCYKGETQESVESRVLPGPRHRH Like above formate the input file contains 1000s of protein sequence. I have to generate thousands of raw file containing only individual protein sequence without any special symbol or gaps. I have developed the code for it in Java but out put is : Cannot open a file followed by cannot find file. Please help me to solve my problem. Regards Vijay Kumar Garg Varanasi Bharat (India) The code is /*Java code to convert FASTA format to a raw format*/ import java.io.*; import java.util.*; import java.util.regex.*; import java.io.FileInputStream; // java package for using regular expression public class Arrayren { public static void main(String args[]) throws IOException { String a[]=new String[1000]; String b[][] =new String[1000][1000]; /*open the id file*/ try { File f = new File ("input.txt"); //opening the text document containing genbank ids FileInputStream fis = new FileInputStream("input.txt"); //Reading the file contents through inputstream BufferedInputStream bis = new BufferedInputStream(fis); // Writing the contents to a buffered stream DataInputStream dis = new DataInputStream(bis); //Method for reading Java Standard data types String inputline; String line; String separator = System.getProperty("line.separator"); // reads a line till next line operator is found int i=0; while ((inputline=dis.readLine()) != null) { i++; a[i]=inputline; a[i]=a[i].replaceAll(separator,""); //replaces unwanted patterns like /n with space a[i]=a[i].trim(); // trims out if any space is available a[i]=a[i]+".txt"; //takes the file name into an array try // to handle run time error /*take the sequence in to an array*/ { BufferedReader in = new BufferedReader (new FileReader(a[i])); String inline = null; int j=0; while((inline=in.readLine()) != null) { j++; b[i][j]=inline; Pattern q=Pattern.compile(">"); //Compiling the regular expression Matcher n=q.matcher(inline); //creates the matcher for the above pattern if(n.find()) { /*appending the comment line*/ b[i][j]=b[i][j].replaceAll(">gi",""); //identify the pattern and replace it with a space b[i][j]=b[i][j].replaceAll("[a-zA-Z]",""); b[i][j]=b[i][j].replaceAll("|",""); b[i][j]=b[i][j].replaceAll("\\d{1,15}",""); b[i][j]=b[i][j].replaceAll(".",""); b[i][j]=b[i][j].replaceAll("_",""); b[i][j]=b[i][j].replaceAll("\\(",""); b[i][j]=b[i][j].replaceAll("\\)",""); } /*printing the sequence in to a text file*/ b[i][j]=b[i][j].replaceAll(separator,""); b[i][j]=b[i][j].trim(); // trims out if any space is available File create = new File(inputline+"R.txt"); try { if(!create.exists()) { create.createNewFile(); // creates a new file } else { System.out.println("file already exists"); } } catch(IOException e) // to catch the exception and print the error if cannot open a file { System.err.println("cannot create a file"); } BufferedWriter outt = new BufferedWriter(new FileWriter(inputline+"R.txt", true)); outt.write(b[i][j]); // printing the contents to a text file outt.close(); // closing the text file System.out.println(b[i][j]); } } catch(Exception e) { System.out.println("cannot open a file"); } } } catch(Exception ex) // catch the exception and prints the error if cannot find file { System.out.println("cannot find file "); } } } If you provide me correct it will be much easier to understand.

    Read the article

  • BioPython: extracting sequence IDs from a Blast output file

    - by Jon
    Hi, I have a BLAST output file in XML format. It is 22 query sequences with 50 hits reported from each sequence. And I want to extract all the 50x22 hits. This is the code I currently have, but it only extracts the 50 hits from the first query. from Bio.Blast import NCBIXM blast_records = NCBIXML.parse(result_handle) blast_record = blast_records.next() save_file = open("/Users/jonbra/Desktop/my_fasta_seq.fasta", 'w') for alignment in blast_record.alignments: for hsp in alignment.hsps: save_file.write('>%s\n' % (alignment.title,)) save_file.close() Somebody have any suggestions as to extract all the hits? I guess I have to use something else than alignments. Hope this was clear. Thanks! Jon

    Read the article

  • Loop over a file and write the next line if a condition is met

    - by 111078384259264152964
    Having a hard time fixing this or finding any good hints about it. I'm trying to loop over one file, modify each line slightly, and then loop over a different file. If the line in the second file starts with the line from the first then the following line in the second file should be written to a third file. !/usr/bin/env python with open('ids.txt', 'rU') as f: with open('seqres.txt', 'rU') as g: for id in f: id=id.lower()[0:4]+'_'+id[4] with open(id + '.fasta', 'w') as h: for line in g: if line.startswith(''+ id): h.write(g.next()) All the correct files appear, but they are empty. Yes, I am sure the if has true cases. :-) "seqres.txt" has lines with an ID number in a certain format, each followed by a line with data. The "ids.txt" has lines with the ID numbers of interest in a different format. I want each line of data with an interesting ID number in its own file. Thanks a million to anyone with a little advice!

    Read the article

  • BASH statements execute alone but return "no such file" in for loop.

    - by reve_etrange
    Another one I can't find an answer for, and it feels like I've gone mad. I have a BASH script using a for loop to run a complex command (many protein sequence alignments) on a lot of files (~5000). The loop produces statements that will execute when given alone (i.e. copy-pasted from the error message to the command prompt), but which return "no such file or directory" inside the loop. Script below; there are actually several more arguments but this includes some representative ones and the file arguments. #!/bin/bash # Pass directory with targets as FASTA sequences as argument. # Arguments to psiblast # Common db=local/db/nr/nr outfile="/mnt/scratch/psi-blast" e=0.001 threads=8 itnum=5 pssm="/mnt/scratch/psi-blast/pssm." pssm_txt="/mnt/scratch/psi-blast/pssm." pseudo=0 pwa_inclusion=0.002 for i in ${1}/* do filename=$(basename $i) "local/ncbi-blast-2.2.23+/bin/psiblast\ -query ${i}\ -db $db\ -out ${outfile}/${filename}.out\ -evalue $e\ -num_threads $threads\ -num_iterations $itnum\ -out_pssm ${pssm}$filename\ -out_ascii_pssm ${pssm_txt}${filename}.txt\ -pseudocount $pseudo\ -inclusion_ethresh $pwa_inclusion" done Running this scripts gives "<scriptname> line <last line before 'done'>: <attempted command> : No such file or directory. If I then paste the attempted command onto the prompt it will run. Each of these commands takes a couple of minutes to run.

    Read the article

  • Faster way to split a string and count characters using R?

    - by chrisamiller
    I'm looking for a faster way to calculate GC content for DNA strings read in from a FASTA file. This boils down to taking a string and counting the number of times that the letter 'G' or 'C' appears. I also want to specify the range of characters to consider. I have a working function that is fairly slow, and it's causing a bottleneck in my code. It looks like this: ## ## count the number of GCs in the characters between start and stop ## gcCount <- function(line, st, sp){ chars = strsplit(as.character(line),"")[[1]] numGC = 0 for(j in st:sp){ ##nested ifs faster than an OR (|) construction if(chars[[j]] == "g"){ numGC <- numGC + 1 }else if(chars[[j]] == "G"){ numGC <- numGC + 1 }else if(chars[[j]] == "c"){ numGC <- numGC + 1 }else if(chars[[j]] == "C"){ numGC <- numGC + 1 } } return(numGC) } Running Rprof gives me the following output: > a = "GCCCAAAATTTTCCGGatttaagcagacataaattcgagg" > Rprof(filename="Rprof.out") > for(i in 1:500000){gcCount(a,1,40)}; > Rprof(NULL) > summaryRprof(filename="Rprof.out") self.time self.pct total.time total.pct "gcCount" 77.36 76.8 100.74 100.0 "==" 18.30 18.2 18.30 18.2 "strsplit" 3.58 3.6 3.64 3.6 "+" 1.14 1.1 1.14 1.1 ":" 0.30 0.3 0.30 0.3 "as.logical" 0.04 0.0 0.04 0.0 "as.character" 0.02 0.0 0.02 0.0 $by.total total.time total.pct self.time self.pct "gcCount" 100.74 100.0 77.36 76.8 "==" 18.30 18.2 18.30 18.2 "strsplit" 3.64 3.6 3.58 3.6 "+" 1.14 1.1 1.14 1.1 ":" 0.30 0.3 0.30 0.3 "as.logical" 0.04 0.0 0.04 0.0 "as.character" 0.02 0.0 0.02 0.0 $sampling.time [1] 100.74 Any advice for making this code faster?

    Read the article

  • How John Got 15x Improvement Without Really Trying

    - by rchrd
    The following article was published on a Sun Microsystems website a number of years ago by John Feo. It is still useful and worth preserving. So I'm republishing it here.  How I Got 15x Improvement Without Really Trying John Feo, Sun Microsystems Taking ten "personal" program codes used in scientific and engineering research, the author was able to get from 2 to 15 times performance improvement easily by applying some simple general optimization techniques. Introduction Scientific research based on computer simulation depends on the simulation for advancement. The research can advance only as fast as the computational codes can execute. The codes' efficiency determines both the rate and quality of results. In the same amount of time, a faster program can generate more results and can carry out a more detailed simulation of physical phenomena than a slower program. Highly optimized programs help science advance quickly and insure that monies supporting scientific research are used as effectively as possible. Scientific computer codes divide into three broad categories: ISV, community, and personal. ISV codes are large, mature production codes developed and sold commercially. The codes improve slowly over time both in methods and capabilities, and they are well tuned for most vendor platforms. Since the codes are mature and complex, there are few opportunities to improve their performance solely through code optimization. Improvements of 10% to 15% are typical. Examples of ISV codes are DYNA3D, Gaussian, and Nastran. Community codes are non-commercial production codes used by a particular research field. Generally, they are developed and distributed by a single academic or research institution with assistance from the community. Most users just run the codes, but some develop new methods and extensions that feed back into the general release. The codes are available on most vendor platforms. Since these codes are younger than ISV codes, there are more opportunities to optimize the source code. Improvements of 50% are not unusual. Examples of community codes are AMBER, CHARM, BLAST, and FASTA. Personal codes are those written by single users or small research groups for their own use. These codes are not distributed, but may be passed from professor-to-student or student-to-student over several years. They form the primordial ocean of applications from which community and ISV codes emerge. Government research grants pay for the development of most personal codes. This paper reports on the nature and performance of this class of codes. Over the last year, I have looked at over two dozen personal codes from more than a dozen research institutions. The codes cover a variety of scientific fields, including astronomy, atmospheric sciences, bioinformatics, biology, chemistry, geology, and physics. The sources range from a few hundred lines to more than ten thousand lines, and are written in Fortran, Fortran 90, C, and C++. For the most part, the codes are modular, documented, and written in a clear, straightforward manner. They do not use complex language features, advanced data structures, programming tricks, or libraries. I had little trouble understanding what the codes did or how data structures were used. Most came with a makefile. Surprisingly, only one of the applications is parallel. All developers have access to parallel machines, so availability is not an issue. Several tried to parallelize their applications, but stopped after encountering difficulties. Lack of education and a perception that parallelism is difficult prevented most from trying. I parallelized several of the codes using OpenMP, and did not judge any of the codes as difficult to parallelize. Even more surprising than the lack of parallelism is the inefficiency of the codes. I was able to get large improvements in performance in a matter of a few days applying simple optimization techniques. Table 1 lists ten representative codes [names and affiliation are omitted to preserve anonymity]. Improvements on one processor range from 2x to 15.5x with a simple average of 4.75x. I did not use sophisticated performance tools or drill deep into the program's execution character as one would do when tuning ISV or community codes. Using only a profiler and source line timers, I identified inefficient sections of code and improved their performance by inspection. The changes were at a high level. I am sure there is another factor of 2 or 3 in each code, and more if the codes are parallelized. The study’s results show that personal scientific codes are running many times slower than they should and that the problem is pervasive. Computational scientists are not sloppy programmers; however, few are trained in the art of computer programming or code optimization. I found that most have a working knowledge of some programming language and standard software engineering practices; but they do not know, or think about, how to make their programs run faster. They simply do not know the standard techniques used to make codes run faster. In fact, they do not even perceive that such techniques exist. The case studies described in this paper show that applying simple, well known techniques can significantly increase the performance of personal codes. It is important that the scientific community and the Government agencies that support scientific research find ways to better educate academic scientific programmers. The inefficiency of their codes is so bad that it is retarding both the quality and progress of scientific research. # cacheperformance redundantoperations loopstructures performanceimprovement 1 x x 15.5 2 x 2.8 3 x x 2.5 4 x 2.1 5 x x 2.0 6 x 5.0 7 x 5.8 8 x 6.3 9 2.2 10 x x 3.3 Table 1 — Area of improvement and performance gains of 10 codes The remainder of the paper is organized as follows: sections 2, 3, and 4 discuss the three most common sources of inefficiencies in the codes studied. These are cache performance, redundant operations, and loop structures. Each section includes several examples. The last section summaries the work and suggests a possible solution to the issues raised. Optimizing cache performance Commodity microprocessor systems use caches to increase memory bandwidth and reduce memory latencies. Typical latencies from processor to L1, L2, local, and remote memory are 3, 10, 50, and 200 cycles, respectively. Moreover, bandwidth falls off dramatically as memory distances increase. Programs that do not use cache effectively run many times slower than programs that do. When optimizing for cache, the biggest performance gains are achieved by accessing data in cache order and reusing data to amortize the overhead of cache misses. Secondary considerations are prefetching, associativity, and replacement; however, the understanding and analysis required to optimize for the latter are probably beyond the capabilities of the non-expert. Much can be gained simply by accessing data in the correct order and maximizing data reuse. 6 out of the 10 codes studied here benefited from such high level optimizations. Array Accesses The most important cache optimization is the most basic: accessing Fortran array elements in column order and C array elements in row order. Four of the ten codes—1, 2, 4, and 10—got it wrong. Compilers will restructure nested loops to optimize cache performance, but may not do so if the loop structure is too complex, or the loop body includes conditionals, complex addressing, or function calls. In code 1, the compiler failed to invert a key loop because of complex addressing do I = 0, 1010, delta_x IM = I - delta_x IP = I + delta_x do J = 5, 995, delta_x JM = J - delta_x JP = J + delta_x T1 = CA1(IP, J) + CA1(I, JP) T2 = CA1(IM, J) + CA1(I, JM) S1 = T1 + T2 - 4 * CA1(I, J) CA(I, J) = CA1(I, J) + D * S1 end do end do In code 2, the culprit is conditionals do I = 1, N do J = 1, N If (IFLAG(I,J) .EQ. 0) then T1 = Value(I, J-1) T2 = Value(I-1, J) T3 = Value(I, J) T4 = Value(I+1, J) T5 = Value(I, J+1) Value(I,J) = 0.25 * (T1 + T2 + T5 + T4) Delta = ABS(T3 - Value(I,J)) If (Delta .GT. MaxDelta) MaxDelta = Delta endif enddo enddo I fixed both programs by inverting the loops by hand. Code 10 has three-dimensional arrays and triply nested loops. The structure of the most computationally intensive loops is too complex to invert automatically or by hand. The only practical solution is to transpose the arrays so that the dimension accessed by the innermost loop is in cache order. The arrays can be transposed at construction or prior to entering a computationally intensive section of code. The former requires all array references to be modified, while the latter is cost effective only if the cost of the transpose is amortized over many accesses. I used the second approach to optimize code 10. Code 5 has four-dimensional arrays and loops are nested four deep. For all of the reasons cited above the compiler is not able to restructure three key loops. Assume C arrays and let the four dimensions of the arrays be i, j, k, and l. In the original code, the index structure of the three loops is L1: for i L2: for i L3: for i for l for l for j for k for j for k for j for k for l So only L3 accesses array elements in cache order. L1 is a very complex loop—much too complex to invert. I brought the loop into cache alignment by transposing the second and fourth dimensions of the arrays. Since the code uses a macro to compute all array indexes, I effected the transpose at construction and changed the macro appropriately. The dimensions of the new arrays are now: i, l, k, and j. L3 is a simple loop and easily inverted. L2 has a loop-carried scalar dependence in k. By promoting the scalar name that carries the dependence to an array, I was able to invert the third and fourth subloops aligning the loop with cache. Code 5 is by far the most difficult of the four codes to optimize for array accesses; but the knowledge required to fix the problems is no more than that required for the other codes. I would judge this code at the limits of, but not beyond, the capabilities of appropriately trained computational scientists. Array Strides When a cache miss occurs, a line (64 bytes) rather than just one word is loaded into the cache. If data is accessed stride 1, than the cost of the miss is amortized over 8 words. Any stride other than one reduces the cost savings. Two of the ten codes studied suffered from non-unit strides. The codes represent two important classes of "strided" codes. Code 1 employs a multi-grid algorithm to reduce time to convergence. The grids are every tenth, fifth, second, and unit element. Since time to convergence is inversely proportional to the distance between elements, coarse grids converge quickly providing good starting values for finer grids. The better starting values further reduce the time to convergence. The downside is that grids of every nth element, n > 1, introduce non-unit strides into the computation. In the original code, much of the savings of the multi-grid algorithm were lost due to this problem. I eliminated the problem by compressing (copying) coarse grids into continuous memory, and rewriting the computation as a function of the compressed grid. On convergence, I copied the final values of the compressed grid back to the original grid. The savings gained from unit stride access of the compressed grid more than paid for the cost of copying. Using compressed grids, the loop from code 1 included in the previous section becomes do j = 1, GZ do i = 1, GZ T1 = CA(i+0, j-1) + CA(i-1, j+0) T4 = CA1(i+1, j+0) + CA1(i+0, j+1) S1 = T1 + T4 - 4 * CA1(i+0, j+0) CA(i+0, j+0) = CA1(i+0, j+0) + DD * S1 enddo enddo where CA and CA1 are compressed arrays of size GZ. Code 7 traverses a list of objects selecting objects for later processing. The labels of the selected objects are stored in an array. The selection step has unit stride, but the processing steps have irregular stride. A fix is to save the parameters of the selected objects in temporary arrays as they are selected, and pass the temporary arrays to the processing functions. The fix is practical if the same parameters are used in selection as in processing, or if processing comprises a series of distinct steps which use overlapping subsets of the parameters. Both conditions are true for code 7, so I achieved significant improvement by copying parameters to temporary arrays during selection. Data reuse In the previous sections, we optimized for spatial locality. It is also important to optimize for temporal locality. Once read, a datum should be used as much as possible before it is forced from cache. Loop fusion and loop unrolling are two techniques that increase temporal locality. Unfortunately, both techniques increase register pressure—as loop bodies become larger, the number of registers required to hold temporary values grows. Once register spilling occurs, any gains evaporate quickly. For multiprocessors with small register sets or small caches, the sweet spot can be very small. In the ten codes presented here, I found no opportunities for loop fusion and only two opportunities for loop unrolling (codes 1 and 3). In code 1, unrolling the outer and inner loop one iteration increases the number of result values computed by the loop body from 1 to 4, do J = 1, GZ-2, 2 do I = 1, GZ-2, 2 T1 = CA1(i+0, j-1) + CA1(i-1, j+0) T2 = CA1(i+1, j-1) + CA1(i+0, j+0) T3 = CA1(i+0, j+0) + CA1(i-1, j+1) T4 = CA1(i+1, j+0) + CA1(i+0, j+1) T5 = CA1(i+2, j+0) + CA1(i+1, j+1) T6 = CA1(i+1, j+1) + CA1(i+0, j+2) T7 = CA1(i+2, j+1) + CA1(i+1, j+2) S1 = T1 + T4 - 4 * CA1(i+0, j+0) S2 = T2 + T5 - 4 * CA1(i+1, j+0) S3 = T3 + T6 - 4 * CA1(i+0, j+1) S4 = T4 + T7 - 4 * CA1(i+1, j+1) CA(i+0, j+0) = CA1(i+0, j+0) + DD * S1 CA(i+1, j+0) = CA1(i+1, j+0) + DD * S2 CA(i+0, j+1) = CA1(i+0, j+1) + DD * S3 CA(i+1, j+1) = CA1(i+1, j+1) + DD * S4 enddo enddo The loop body executes 12 reads, whereas as the rolled loop shown in the previous section executes 20 reads to compute the same four values. In code 3, two loops are unrolled 8 times and one loop is unrolled 4 times. Here is the before for (k = 0; k < NK[u]; k++) { sum = 0.0; for (y = 0; y < NY; y++) { sum += W[y][u][k] * delta[y]; } backprop[i++]=sum; } and after code for (k = 0; k < KK - 8; k+=8) { sum0 = 0.0; sum1 = 0.0; sum2 = 0.0; sum3 = 0.0; sum4 = 0.0; sum5 = 0.0; sum6 = 0.0; sum7 = 0.0; for (y = 0; y < NY; y++) { sum0 += W[y][0][k+0] * delta[y]; sum1 += W[y][0][k+1] * delta[y]; sum2 += W[y][0][k+2] * delta[y]; sum3 += W[y][0][k+3] * delta[y]; sum4 += W[y][0][k+4] * delta[y]; sum5 += W[y][0][k+5] * delta[y]; sum6 += W[y][0][k+6] * delta[y]; sum7 += W[y][0][k+7] * delta[y]; } backprop[k+0] = sum0; backprop[k+1] = sum1; backprop[k+2] = sum2; backprop[k+3] = sum3; backprop[k+4] = sum4; backprop[k+5] = sum5; backprop[k+6] = sum6; backprop[k+7] = sum7; } for one of the loops unrolled 8 times. Optimizing for temporal locality is the most difficult optimization considered in this paper. The concepts are not difficult, but the sweet spot is small. Identifying where the program can benefit from loop unrolling or loop fusion is not trivial. Moreover, it takes some effort to get it right. Still, educating scientific programmers about temporal locality and teaching them how to optimize for it will pay dividends. Reducing instruction count Execution time is a function of instruction count. Reduce the count and you usually reduce the time. The best solution is to use a more efficient algorithm; that is, an algorithm whose order of complexity is smaller, that converges quicker, or is more accurate. Optimizing source code without changing the algorithm yields smaller, but still significant, gains. This paper considers only the latter because the intent is to study how much better codes can run if written by programmers schooled in basic code optimization techniques. The ten codes studied benefited from three types of "instruction reducing" optimizations. The two most prevalent were hoisting invariant memory and data operations out of inner loops. The third was eliminating unnecessary data copying. The nature of these inefficiencies is language dependent. Memory operations The semantics of C make it difficult for the compiler to determine all the invariant memory operations in a loop. The problem is particularly acute for loops in functions since the compiler may not know the values of the function's parameters at every call site when compiling the function. Most compilers support pragmas to help resolve ambiguities; however, these pragmas are not comprehensive and there is no standard syntax. To guarantee that invariant memory operations are not executed repetitively, the user has little choice but to hoist the operations by hand. The problem is not as severe in Fortran programs because in the absence of equivalence statements, it is a violation of the language's semantics for two names to share memory. Codes 3 and 5 are C programs. In both cases, the compiler did not hoist all invariant memory operations from inner loops. Consider the following loop from code 3 for (y = 0; y < NY; y++) { i = 0; for (u = 0; u < NU; u++) { for (k = 0; k < NK[u]; k++) { dW[y][u][k] += delta[y] * I1[i++]; } } } Since dW[y][u] can point to the same memory space as delta for one or more values of y and u, assignment to dW[y][u][k] may change the value of delta[y]. In reality, dW and delta do not overlap in memory, so I rewrote the loop as for (y = 0; y < NY; y++) { i = 0; Dy = delta[y]; for (u = 0; u < NU; u++) { for (k = 0; k < NK[u]; k++) { dW[y][u][k] += Dy * I1[i++]; } } } Failure to hoist invariant memory operations may be due to complex address calculations. If the compiler can not determine that the address calculation is invariant, then it can hoist neither the calculation nor the associated memory operations. As noted above, code 5 uses a macro to address four-dimensional arrays #define MAT4D(a,q,i,j,k) (double *)((a)->data + (q)*(a)->strides[0] + (i)*(a)->strides[3] + (j)*(a)->strides[2] + (k)*(a)->strides[1]) The macro is too complex for the compiler to understand and so, it does not identify any subexpressions as loop invariant. The simplest way to eliminate the address calculation from the innermost loop (over i) is to define a0 = MAT4D(a,q,0,j,k) before the loop and then replace all instances of *MAT4D(a,q,i,j,k) in the loop with a0[i] A similar problem appears in code 6, a Fortran program. The key loop in this program is do n1 = 1, nh nx1 = (n1 - 1) / nz + 1 nz1 = n1 - nz * (nx1 - 1) do n2 = 1, nh nx2 = (n2 - 1) / nz + 1 nz2 = n2 - nz * (nx2 - 1) ndx = nx2 - nx1 ndy = nz2 - nz1 gxx = grn(1,ndx,ndy) gyy = grn(2,ndx,ndy) gxy = grn(3,ndx,ndy) balance(n1,1) = balance(n1,1) + (force(n2,1) * gxx + force(n2,2) * gxy) * h1 balance(n1,2) = balance(n1,2) + (force(n2,1) * gxy + force(n2,2) * gyy)*h1 end do end do The programmer has written this loop well—there are no loop invariant operations with respect to n1 and n2. However, the loop resides within an iterative loop over time and the index calculations are independent with respect to time. Trading space for time, I precomputed the index values prior to the entering the time loop and stored the values in two arrays. I then replaced the index calculations with reads of the arrays. Data operations Ways to reduce data operations can appear in many forms. Implementing a more efficient algorithm produces the biggest gains. The closest I came to an algorithm change was in code 4. This code computes the inner product of K-vectors A(i) and B(j), 0 = i < N, 0 = j < M, for most values of i and j. Since the program computes most of the NM possible inner products, it is more efficient to compute all the inner products in one triply-nested loop rather than one at a time when needed. The savings accrue from reading A(i) once for all B(j) vectors and from loop unrolling. for (i = 0; i < N; i+=8) { for (j = 0; j < M; j++) { sum0 = 0.0; sum1 = 0.0; sum2 = 0.0; sum3 = 0.0; sum4 = 0.0; sum5 = 0.0; sum6 = 0.0; sum7 = 0.0; for (k = 0; k < K; k++) { sum0 += A[i+0][k] * B[j][k]; sum1 += A[i+1][k] * B[j][k]; sum2 += A[i+2][k] * B[j][k]; sum3 += A[i+3][k] * B[j][k]; sum4 += A[i+4][k] * B[j][k]; sum5 += A[i+5][k] * B[j][k]; sum6 += A[i+6][k] * B[j][k]; sum7 += A[i+7][k] * B[j][k]; } C[i+0][j] = sum0; C[i+1][j] = sum1; C[i+2][j] = sum2; C[i+3][j] = sum3; C[i+4][j] = sum4; C[i+5][j] = sum5; C[i+6][j] = sum6; C[i+7][j] = sum7; }} This change requires knowledge of a typical run; i.e., that most inner products are computed. The reasons for the change, however, derive from basic optimization concepts. It is the type of change easily made at development time by a knowledgeable programmer. In code 5, we have the data version of the index optimization in code 6. Here a very expensive computation is a function of the loop indices and so cannot be hoisted out of the loop; however, the computation is invariant with respect to an outer iterative loop over time. We can compute its value for each iteration of the computation loop prior to entering the time loop and save the values in an array. The increase in memory required to store the values is small in comparison to the large savings in time. The main loop in Code 8 is doubly nested. The inner loop includes a series of guarded computations; some are a function of the inner loop index but not the outer loop index while others are a function of the outer loop index but not the inner loop index for (j = 0; j < N; j++) { for (i = 0; i < M; i++) { r = i * hrmax; R = A[j]; temp = (PRM[3] == 0.0) ? 1.0 : pow(r, PRM[3]); high = temp * kcoeff * B[j] * PRM[2] * PRM[4]; low = high * PRM[6] * PRM[6] / (1.0 + pow(PRM[4] * PRM[6], 2.0)); kap = (R > PRM[6]) ? high * R * R / (1.0 + pow(PRM[4]*r, 2.0) : low * pow(R/PRM[6], PRM[5]); < rest of loop omitted > }} Note that the value of temp is invariant to j. Thus, we can hoist the computation for temp out of the loop and save its values in an array. for (i = 0; i < M; i++) { r = i * hrmax; TEMP[i] = pow(r, PRM[3]); } [N.B. – the case for PRM[3] = 0 is omitted and will be reintroduced later.] We now hoist out of the inner loop the computations invariant to i. Since the conditional guarding the value of kap is invariant to i, it behooves us to hoist the computation out of the inner loop, thereby executing the guard once rather than M times. The final version of the code is for (j = 0; j < N; j++) { R = rig[j] / 1000.; tmp1 = kcoeff * par[2] * beta[j] * par[4]; tmp2 = 1.0 + (par[4] * par[4] * par[6] * par[6]); tmp3 = 1.0 + (par[4] * par[4] * R * R); tmp4 = par[6] * par[6] / tmp2; tmp5 = R * R / tmp3; tmp6 = pow(R / par[6], par[5]); if ((par[3] == 0.0) && (R > par[6])) { for (i = 1; i <= imax1; i++) KAP[i] = tmp1 * tmp5; } else if ((par[3] == 0.0) && (R <= par[6])) { for (i = 1; i <= imax1; i++) KAP[i] = tmp1 * tmp4 * tmp6; } else if ((par[3] != 0.0) && (R > par[6])) { for (i = 1; i <= imax1; i++) KAP[i] = tmp1 * TEMP[i] * tmp5; } else if ((par[3] != 0.0) && (R <= par[6])) { for (i = 1; i <= imax1; i++) KAP[i] = tmp1 * TEMP[i] * tmp4 * tmp6; } for (i = 0; i < M; i++) { kap = KAP[i]; r = i * hrmax; < rest of loop omitted > } } Maybe not the prettiest piece of code, but certainly much more efficient than the original loop, Copy operations Several programs unnecessarily copy data from one data structure to another. This problem occurs in both Fortran and C programs, although it manifests itself differently in the two languages. Code 1 declares two arrays—one for old values and one for new values. At the end of each iteration, the array of new values is copied to the array of old values to reset the data structures for the next iteration. This problem occurs in Fortran programs not included in this study and in both Fortran 77 and Fortran 90 code. Introducing pointers to the arrays and swapping pointer values is an obvious way to eliminate the copying; but pointers is not a feature that many Fortran programmers know well or are comfortable using. An easy solution not involving pointers is to extend the dimension of the value array by 1 and use the last dimension to differentiate between arrays at different times. For example, if the data space is N x N, declare the array (N, N, 2). Then store the problem’s initial values in (_, _, 2) and define the scalar names new = 2 and old = 1. At the start of each iteration, swap old and new to reset the arrays. The old–new copy problem did not appear in any C program. In programs that had new and old values, the code swapped pointers to reset data structures. Where unnecessary coping did occur is in structure assignment and parameter passing. Structures in C are handled much like scalars. Assignment causes the data space of the right-hand name to be copied to the data space of the left-hand name. Similarly, when a structure is passed to a function, the data space of the actual parameter is copied to the data space of the formal parameter. If the structure is large and the assignment or function call is in an inner loop, then copying costs can grow quite large. While none of the ten programs considered here manifested this problem, it did occur in programs not included in the study. A simple fix is always to refer to structures via pointers. Optimizing loop structures Since scientific programs spend almost all their time in loops, efficient loops are the key to good performance. Conditionals, function calls, little instruction level parallelism, and large numbers of temporary values make it difficult for the compiler to generate tightly packed, highly efficient code. Conditionals and function calls introduce jumps that disrupt code flow. Users should eliminate or isolate conditionls to their own loops as much as possible. Often logical expressions can be substituted for if-then-else statements. For example, code 2 includes the following snippet MaxDelta = 0.0 do J = 1, N do I = 1, M < code omitted > Delta = abs(OldValue ? NewValue) if (Delta > MaxDelta) MaxDelta = Delta enddo enddo if (MaxDelta .gt. 0.001) goto 200 Since the only use of MaxDelta is to control the jump to 200 and all that matters is whether or not it is greater than 0.001, I made MaxDelta a boolean and rewrote the snippet as MaxDelta = .false. do J = 1, N do I = 1, M < code omitted > Delta = abs(OldValue ? NewValue) MaxDelta = MaxDelta .or. (Delta .gt. 0.001) enddo enddo if (MaxDelta) goto 200 thereby, eliminating the conditional expression from the inner loop. A microprocessor can execute many instructions per instruction cycle. Typically, it can execute one or more memory, floating point, integer, and jump operations. To be executed simultaneously, the operations must be independent. Thick loops tend to have more instruction level parallelism than thin loops. Moreover, they reduce memory traffice by maximizing data reuse. Loop unrolling and loop fusion are two techniques to increase the size of loop bodies. Several of the codes studied benefitted from loop unrolling, but none benefitted from loop fusion. This observation is not too surpising since it is the general tendency of programmers to write thick loops. As loops become thicker, the number of temporary values grows, increasing register pressure. If registers spill, then memory traffic increases and code flow is disrupted. A thick loop with many temporary values may execute slower than an equivalent series of thin loops. The biggest gain will be achieved if the thick loop can be split into a series of independent loops eliminating the need to write and read temporary arrays. I found such an occasion in code 10 where I split the loop do i = 1, n do j = 1, m A24(j,i)= S24(j,i) * T24(j,i) + S25(j,i) * U25(j,i) B24(j,i)= S24(j,i) * T25(j,i) + S25(j,i) * U24(j,i) A25(j,i)= S24(j,i) * C24(j,i) + S25(j,i) * V24(j,i) B25(j,i)= S24(j,i) * U25(j,i) + S25(j,i) * V25(j,i) C24(j,i)= S26(j,i) * T26(j,i) + S27(j,i) * U26(j,i) D24(j,i)= S26(j,i) * T27(j,i) + S27(j,i) * V26(j,i) C25(j,i)= S27(j,i) * S28(j,i) + S26(j,i) * U28(j,i) D25(j,i)= S27(j,i) * T28(j,i) + S26(j,i) * V28(j,i) end do end do into two disjoint loops do i = 1, n do j = 1, m A24(j,i)= S24(j,i) * T24(j,i) + S25(j,i) * U25(j,i) B24(j,i)= S24(j,i) * T25(j,i) + S25(j,i) * U24(j,i) A25(j,i)= S24(j,i) * C24(j,i) + S25(j,i) * V24(j,i) B25(j,i)= S24(j,i) * U25(j,i) + S25(j,i) * V25(j,i) end do end do do i = 1, n do j = 1, m C24(j,i)= S26(j,i) * T26(j,i) + S27(j,i) * U26(j,i) D24(j,i)= S26(j,i) * T27(j,i) + S27(j,i) * V26(j,i) C25(j,i)= S27(j,i) * S28(j,i) + S26(j,i) * U28(j,i) D25(j,i)= S27(j,i) * T28(j,i) + S26(j,i) * V28(j,i) end do end do Conclusions Over the course of the last year, I have had the opportunity to work with over two dozen academic scientific programmers at leading research universities. Their research interests span a broad range of scientific fields. Except for two programs that relied almost exclusively on library routines (matrix multiply and fast Fourier transform), I was able to improve significantly the single processor performance of all codes. Improvements range from 2x to 15.5x with a simple average of 4.75x. Changes to the source code were at a very high level. I did not use sophisticated techniques or programming tools to discover inefficiencies or effect the changes. Only one code was parallel despite the availability of parallel systems to all developers. Clearly, we have a problem—personal scientific research codes are highly inefficient and not running parallel. The developers are unaware of simple optimization techniques to make programs run faster. They lack education in the art of code optimization and parallel programming. I do not believe we can fix the problem by publishing additional books or training manuals. To date, the developers in questions have not studied the books or manual available, and are unlikely to do so in the future. Short courses are a possible solution, but I believe they are too concentrated to be much use. The general concepts can be taught in a three or four day course, but that is not enough time for students to practice what they learn and acquire the experience to apply and extend the concepts to their codes. Practice is the key to becoming proficient at optimization. I recommend that graduate students be required to take a semester length course in optimization and parallel programming. We would never give someone access to state-of-the-art scientific equipment costing hundreds of thousands of dollars without first requiring them to demonstrate that they know how to use the equipment. Yet the criterion for time on state-of-the-art supercomputers is at most an interesting project. Requestors are never asked to demonstrate that they know how to use the system, or can use the system effectively. A semester course would teach them the required skills. Government agencies that fund academic scientific research pay for most of the computer systems supporting scientific research as well as the development of most personal scientific codes. These agencies should require graduate schools to offer a course in optimization and parallel programming as a requirement for funding. About the Author John Feo received his Ph.D. in Computer Science from The University of Texas at Austin in 1986. After graduate school, Dr. Feo worked at Lawrence Livermore National Laboratory where he was the Group Leader of the Computer Research Group and principal investigator of the Sisal Language Project. In 1997, Dr. Feo joined Tera Computer Company where he was project manager for the MTA, and oversaw the programming and evaluation of the MTA at the San Diego Supercomputer Center. In 2000, Dr. Feo joined Sun Microsystems as an HPC application specialist. He works with university research groups to optimize and parallelize scientific codes. Dr. Feo has published over two dozen research articles in the areas of parallel parallel programming, parallel programming languages, and application performance.

    Read the article

1