Linux 3.16 I/O Scheduler

Linux 3.16 kernel IO scheduler benchmarks of Deadline CFQ etc. Benchmarks by Michael Larabel for a future article on Phoronix.

Compare your own system(s) to this result file with the Phoronix Test Suite by running the command: phoronix-test-suite benchmark 1406238-KH-LINUX316I30
Jump To Table - Results

View

Do Not Show Noisy Results
Do Not Show Results With Incomplete Data
Do Not Show Results With Little Change/Spread
List Notable Results

Limit displaying results to tests within:

CPU Massive 3 Tests
Disk Test Suite 4 Tests
Common Kernel Benchmarks 2 Tests
Server 2 Tests

Statistics

Show Overall Harmonic Mean(s)
Show Overall Geometric Mean
Show Geometric Means Per-Suite/Category
Show Wins / Losses Counts (Pie Chart)
Normalize Results
Remove Outliers Before Calculating Averages

Graph Settings

Force Line Graphs Where Applicable
Convert To Scalar Where Applicable
Prefer Vertical Bar Graphs

Multi-Way Comparison

Condense Multi-Option Tests Into Single Result Graphs

Table

Show Detailed System Result Table

Run Management

Highlight
Result
Hide
Result
Result
Identifier
View Logs
Performance Per
Dollar
Date
Run
  Test
  Duration
Deadline
June 22 2014
 
Noop
June 22 2014
 
CFQ
June 22 2014
 
Invert Hiding All Results Option
 

Only show results where is faster than
Only show results matching title/arguments (delimit multiple options with a comma):
Do not show results matching title/arguments (delimit multiple options with a comma):


Linux 3.16 I/O SchedulerOpenBenchmarking.orgPhoronix Test SuiteIntel Core i7-4790K @ 4.40GHz (8 Cores)ASRock Z97 Extreme6Intel 4th Gen Core DRAM16384MB128GB Crucial_CT128MX1Intel HD 4600Intel Xeon E3-1200 v3/4thIntel ConnectionUbuntu 14.043.16.0-999-generic (x86_64)Unity 7.2.1intel 2.99.912GCC 4.8.2ext41024x768ProcessorMotherboardChipsetMemoryDiskGraphicsAudioNetworkOSKernelDesktopDisplay DriverCompilerFile-SystemScreen ResolutionLinux 3.16 I/O Scheduler BenchmarksSystem Logs- --build=x86_64-linux-gnu --disable-browser-plugin --disable-libmudflap --disable-werror --enable-checking=release --enable-clocale=gnu --enable-gnu-unique-object --enable-gtk-cairo --enable-java-awt=gtk --enable-java-home --enable-languages=c,c++,java,go,d,fortran,objc,obj-c++ --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-nls --enable-objc-gc --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-arch-directory=amd64 --with-multilib-list=m32,m64,mx32 --with-tune=generic -v - Deadline: DEADLINE / data=ordered,errors=remount-ro,relatime,rw - Noop: NOOP / data=ordered,errors=remount-ro,relatime,rw - CFQ: CFQ / data=ordered,errors=remount-ro,relatime,rw - Scaling Governor: intel_pstate performance- Deadline: Disk Scheduler: DEADLINE. Python 2.7.6.- Noop: Disk Scheduler: NOOP. Python 2.7.6.- CFQ: Disk Scheduler: CFQ. Python 2.7.6.

DeadlineNoopCFQResult OverviewPhoronix Test Suite100%111%123%134%Flexible IO TesterFS-MarkFS-MarkFS-MarkBlogBenchThreaded I/O TesterCompile BenchApache BenchmarkCompile BenchPostMarkCompile BenchI.I.F.S.A.P5.F.1.S.4.T1.F.1.S4.F.3.S.D.1.SWriteRand Write - 32MB - 16Initial CreateS.W.P.SCompileD.T.PRead Compiled Tree

Linux 3.16 I/O Schedulerfio: Intel IOMeter File Server Access Patternfs-mark: 1000 Files, 1MB Sizefs-mark: 5000 Files, 1MB Size, 4 Threadsfs-mark: 4000 Files, 32 Sub Dirs, 1MB Sizeblogbench: Writetiobench: Rand Write - 32MB - 16compilebench: Compilecompilebench: Initial Createcompilebench: Read Compiled Treepostmark: Disk Transaction Performanceapache: Static Web Page ServingDeadlineNoopCFQ150.6816.7339.8018.7226770.57266.90173.782685.8558135520.82207.8613.4233.4016.4823850.55271.00168.152667.5758135419.20219.7512.9327.4014.5727370.59271.82164.892669.9358595544.94OpenBenchmarking.org

Flexible IO Tester

OpenBenchmarking.orgSeconds (Run Time), Fewer Is BetterFlexible IO Tester 2.1.9Test: Intel IOMeter File Server Access PatternDeadlineNoopCFQ50100150200250SE +/- 4.92, N = 6SE +/- 14.43, N = 6SE +/- 13.99, N = 6150.68207.86219.751. (CC) gcc options: -rdynamic -std=gnu99 -O3 -ffast-math -include -libverbs -lrt -laio -lz -lm -lpthread -ldl
OpenBenchmarking.orgSeconds (Run Time), Fewer Is BetterFlexible IO Tester 2.1.9Test: Intel IOMeter File Server Access PatternDeadlineNoopCFQ4080120160200Min: 129.37 / Avg: 150.68 / Max: 164.4Min: 155.88 / Avg: 207.86 / Max: 238.67Min: 153.4 / Avg: 219.75 / Max: 242.181. (CC) gcc options: -rdynamic -std=gnu99 -O3 -ffast-math -include -libverbs -lrt -laio -lz -lm -lpthread -ldl

FS-Mark

OpenBenchmarking.orgFiles/s, More Is BetterFS-Mark 3.3Test: 1000 Files, 1MB SizeDeadlineNoopCFQ48121620SE +/- 0.70, N = 6SE +/- 0.53, N = 6SE +/- 0.39, N = 616.7313.4212.931. (CC) gcc options: -static
OpenBenchmarking.orgFiles/s, More Is BetterFS-Mark 3.3Test: 1000 Files, 1MB SizeDeadlineNoopCFQ48121620Min: 15.3 / Avg: 16.73 / Max: 20Min: 11.1 / Avg: 13.42 / Max: 15Min: 11.3 / Avg: 12.93 / Max: 13.91. (CC) gcc options: -static

OpenBenchmarking.orgFiles/s, More Is BetterFS-Mark 3.3Test: 5000 Files, 1MB Size, 4 ThreadsDeadlineNoopCFQ918273645SE +/- 1.28, N = 6SE +/- 1.54, N = 6SE +/- 1.17, N = 639.8033.4027.401. (CC) gcc options: -static
OpenBenchmarking.orgFiles/s, More Is BetterFS-Mark 3.3Test: 5000 Files, 1MB Size, 4 ThreadsDeadlineNoopCFQ816243240Min: 36 / Avg: 39.8 / Max: 44Min: 28.8 / Avg: 33.4 / Max: 39.6Min: 23.2 / Avg: 27.4 / Max: 31.61. (CC) gcc options: -static

OpenBenchmarking.orgFiles/s, More Is BetterFS-Mark 3.3Test: 4000 Files, 32 Sub Dirs, 1MB SizeDeadlineNoopCFQ510152025SE +/- 0.75, N = 6SE +/- 0.75, N = 6SE +/- 0.30, N = 618.7216.4814.571. (CC) gcc options: -static
OpenBenchmarking.orgFiles/s, More Is BetterFS-Mark 3.3Test: 4000 Files, 32 Sub Dirs, 1MB SizeDeadlineNoopCFQ510152025Min: 17.3 / Avg: 18.72 / Max: 22.3Min: 14.9 / Avg: 16.48 / Max: 19.8Min: 14 / Avg: 14.57 / Max: 161. (CC) gcc options: -static

BlogBench

OpenBenchmarking.orgFinal Score, More Is BetterBlogBench 1.0Test: WriteDeadlineNoopCFQ6001200180024003000SE +/- 42.07, N = 6SE +/- 88.43, N = 6SE +/- 7.22, N = 32677238527371. (CC) gcc options: -O2 -pthread
OpenBenchmarking.orgFinal Score, More Is BetterBlogBench 1.0Test: WriteDeadlineNoopCFQ5001000150020002500Min: 2523 / Avg: 2676.5 / Max: 2837Min: 2229 / Avg: 2385.17 / Max: 2791Min: 2725 / Avg: 2737.33 / Max: 27501. (CC) gcc options: -O2 -pthread

Threaded I/O Tester

OpenBenchmarking.orgMB/s, More Is BetterThreaded I/O Tester 0.3.3Test: Random Write - Size Per Thread: 32MB - Thread Count: 16DeadlineNoopCFQ0.13280.26560.39840.53120.664SE +/- 0.01, N = 3SE +/- 0.00, N = 3SE +/- 0.01, N = 30.570.550.59
OpenBenchmarking.orgMB/s, More Is BetterThreaded I/O Tester 0.3.3Test: Random Write - Size Per Thread: 32MB - Thread Count: 16DeadlineNoopCFQ246810Min: 0.56 / Avg: 0.57 / Max: 0.58Min: 0.55 / Avg: 0.55 / Max: 0.56Min: 0.57 / Avg: 0.59 / Max: 0.61

Compile Bench

Compilebench tries to age a filesystem by simulating some of the disk IO common in creating, compiling, patching, stating and reading kernel trees. It indirectly measures how well filesystems can maintain directory locality as the disk fills up and directories age. This current test is setup to use the makej mode with 10 initial directories Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMB/s, More Is BetterCompile Bench 0.6Test: CompileDeadlineNoopCFQ60120180240300SE +/- 1.70, N = 3SE +/- 2.85, N = 3SE +/- 4.86, N = 3266.90271.00271.82
OpenBenchmarking.orgMB/s, More Is BetterCompile Bench 0.6Test: CompileDeadlineNoopCFQ50100150200250Min: 263.96 / Avg: 266.9 / Max: 269.84Min: 265.31 / Avg: 271 / Max: 273.96Min: 263.96 / Avg: 271.82 / Max: 280.69

OpenBenchmarking.orgMB/s, More Is BetterCompile Bench 0.6Test: Initial CreateDeadlineNoopCFQ4080120160200SE +/- 26.59, N = 3SE +/- 21.55, N = 3SE +/- 22.30, N = 3173.78168.15164.89
OpenBenchmarking.orgMB/s, More Is BetterCompile Bench 0.6Test: Initial CreateDeadlineNoopCFQ306090120150Min: 120.7 / Avg: 173.78 / Max: 203.06Min: 125.42 / Avg: 168.15 / Max: 194.35Min: 120.56 / Avg: 164.89 / Max: 191.38

OpenBenchmarking.orgMB/s, More Is BetterCompile Bench 0.6Test: Read Compiled TreeDeadlineNoopCFQ6001200180024003000SE +/- 9.69, N = 3SE +/- 14.20, N = 3SE +/- 16.30, N = 32685.852667.572669.93
OpenBenchmarking.orgMB/s, More Is BetterCompile Bench 0.6Test: Read Compiled TreeDeadlineNoopCFQ5001000150020002500Min: 2667.1 / Avg: 2685.85 / Max: 2699.47Min: 2643.21 / Avg: 2667.57 / Max: 2692.4Min: 2643.21 / Avg: 2669.93 / Max: 2699.47

PostMark

This is a test of NetApp's PostMark benchmark designed to simulate small-file testing similar to the tasks endured by web and mail servers. This test profile will set PostMark to perform 25,000 transactions with 500 files simultaneously with the file sizes ranging between 5 and 512 kilobytes. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgTPS, More Is BetterPostMark 1.51Disk Transaction PerformanceDeadlineNoopCFQ13002600390052006500SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 46.33, N = 35813581358591. (CC) gcc options: -O3
OpenBenchmarking.orgTPS, More Is BetterPostMark 1.51Disk Transaction PerformanceDeadlineNoopCFQ10002000300040005000Min: 5813 / Avg: 5813 / Max: 5813Min: 5813 / Avg: 5813 / Max: 5813Min: 5813 / Avg: 5859.33 / Max: 59521. (CC) gcc options: -O3

Apache Benchmark

This is a test of ab, which is the Apache benchmark program. This test profile measures how many requests per second a given system can sustain when carrying out 1,000,000 requests with 100 requests being carried out concurrently. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgRequests Per Second, More Is BetterApache Benchmark 2.4.7Static Web Page ServingDeadlineNoopCFQ12002400360048006000SE +/- 51.89, N = 3SE +/- 36.19, N = 3SE +/- 46.97, N = 35520.825419.205544.941. (CC) gcc options: -shared -fPIC -pthread -O3 -march=native
OpenBenchmarking.orgRequests Per Second, More Is BetterApache Benchmark 2.4.7Static Web Page ServingDeadlineNoopCFQ10002000300040005000Min: 5417.04 / Avg: 5520.82 / Max: 5573.85Min: 5346.9 / Avg: 5419.2 / Max: 5458.49Min: 5461.46 / Avg: 5544.94 / Max: 5623.991. (CC) gcc options: -shared -fPIC -pthread -O3 -march=native