summaryrefslogtreecommitdiff
path: root/Microbenchmarks/STREAM
AgeCommit message (Collapse)AuthorLines
9 daysSTREAM: Re-enable -fltoHEADmainBirte Kristina Friesel-1/+1
SDK 2024.1.0 shipped a compiler bug. With 2024.2.0, it is working again.
2024-08-23Make STREAM Microbenchmark work with SDK 2024.1.0Birte Kristina Friesel-1/+1
For whatever reason, -flto causes a "Heap Full" error even in an otherwise empty program: * thread #1, name = 'DPUthread0', stop reason = fault 1 (Heap Full) * frame #0: 0x80000350 dpu_code`mem_alloc_nolock(size=64) at alloc.c:32:5 frame #1: 0x80000170 dpu_code`main_kernel1 [inlined] mem_alloc(size=64) at alloc.c:52:21 frame #2: 0x80000168 dpu_code`main_kernel1 at add.c:73 frame #3: 0x80000308 dpu_code`main at add.c:42:12 frame #4: 0x80000050 dpu_code`__bootstrap at crt0.c:36:5 I do not know why this is the case.
2024-05-13Print more info, set localeMarcel Köppen-1/+4
2024-05-13Flush stdout after printing iteration resultsMarcel Köppen-0/+1
2024-05-13Test different parameter setMarcel Köppen-3/+3
2024-05-13Parameterize iteration count and timeoutMarcel Köppen-1/+4
2024-05-13Treat unroll like the other parametersMarcel Köppen-7/+6
2024-05-13Print hostname and compiler versionsMarcel Köppen-1/+7
2024-05-13Make the linker happy by switching HOST_SOURCES and HOST_FLAGSMarcel Köppen-1/+1
2024-05-13Add depenceny on bin directoryMarcel Köppen-2/+2
2024-05-13past derf has been naughty and not committed changes in timeBirte Kristina Friesel-10/+12
2024-03-04STREAM: support SERIAL (default) and PUSH (new) transfersBirte Kristina Friesel-9/+46
2024-03-04STREAM: Include date in output file nameBirte Kristina Friesel-2/+2
2024-02-29STREAM: adjust run-rank BL rangeBirte Kristina Friesel-3/+3
2024-02-23STREAM: shell scripting is hard, let's go shoppingBirte Kristina Friesel-0/+1
2024-02-23STREAM: no alloc/load overhead for nowBirte Kristina Friesel-2/+2
2024-02-22STREAM: nitsBirte Kristina Friesel-37/+3
2024-02-22STREAM: Add idle and stress versions of run-rankBirte Kristina Friesel-6/+24
2024-02-22STREAM run-rank.sh: -vBirte Kristina Friesel-10/+10
2024-02-22STREAM: Use nano- rather than microsecond precision internallyBirte Kristina Friesel-41/+77
2024-02-21STREAM output: Add n_elements_per_dpuBirte Kristina Friesel-2/+2
2023-11-27STREAM: Add alloc/load overhead, n_ranks, and n_elementsBirte Kristina Friesel-61/+169
2023-05-25re-add (and, in some cases, fix) -x supportDaniel Friesel-2/+5
2023-05-24run benchmarks with timeout of 30 minutes per configurationDaniel Friesel-1/+1
2023-05-11STREAM: distinguish between COPY and COPYWDaniel Friesel-3/+5
2023-05-11STREAM: remove timing outputDaniel Friesel-11/+0
2023-05-11STREAM: Do not include performance data readout in performance data calculationDaniel Friesel-4/+7
2023-05-11STREAM: some quality of life improvementsDaniel Friesel-7/+15
2023-05-11rn.sh: log date and revisionDaniel Friesel-3/+7
2023-05-10derpDaniel Friesel-2/+2
2023-05-10STREAM: Use upstream MB/s calculation methodDaniel Friesel-1/+7
2023-05-10re-order run.shDaniel Friesel-1/+1
2023-05-10STREAM: dfatool compat; variable data typeDaniel Friesel-94/+107
2023-05-10support: chmod 644Daniel Friesel-0/+0
2023-04-24microbenchmarks: create "profile" directory if it does not exist; set -eDerf Null-0/+3
2021-06-16PrIM -- first commitJuan Gomez Luna-0/+987