102 Commits

Author SHA1 Message Date
Julian
cec6d5cbf2 Merge pull request #3 from Tobi29/aarch64
Add explicit support for AArch64
2020-10-26 10:14:45 +01:00
Git out :V
d18b4d2004 Add explicit support for AArch64
Specifies register shape and suffix for floating point and vector
registers. The former cannot be benchmarked without and the latter would
require adding the required suffixes manually in the instruction
operands. Doing both allows using them in the same manner as on x86.

Additionally there are two small changes affecting all architectures:

Allow the 'w' constraint code, which is used for vector registers on
aarch64.

Always specify a clobber for the flags register as many instructions one
might want to benchmark modify it.
2020-10-20 17:43:14 +02:00
JanLJL
9143ac609e Merge branch 'master' of github.com:RRZE-HPC/asmbench 2020-06-08 16:08:14 +02:00
JanLJL
50d585c758 adjusted for non latency mode 2020-06-08 16:06:07 +02:00
Julian Hammer
ba91c9ff02 Revert "fixed up to work with latest kerncraft"
This reverts commit 2ccfb0c9ea.
2020-05-27 14:39:02 +02:00
Julian Hammer
4447e8eec9 Merge branch 'master' of github.com:RRZE-HPC/asmbench 2020-05-27 14:21:37 +02:00
Jan
966d91531f typos 2020-05-27 14:11:01 +02:00
Jan
6e5b808bd7 fixed typos 2020-05-27 14:09:23 +02:00
Julian Hammer
9c511f9ddf Revert "fixed up to work with latest kerncraft"
This reverts commit 2ccfb0c9ea.
2020-05-27 14:09:13 +02:00
Julian Hammer
2ccfb0c9ea fixed up to work with latest kerncraft 2020-02-14 16:57:29 +01:00
Julian Hammer
201227563e throughput benchmark will work if code can not be serialized 2020-01-15 12:46:37 +01:00
Julian
8fe6aa2097 Update README.rst 2020-01-14 10:27:47 +01:00
Julian Hammer
6f62086187 now using max frequency! new version 2020-01-10 15:26:58 +01:00
Julian Hammer
c739acab7e fixed #2. updated help text 2019-10-16 15:00:30 +02:00
Julian Hammer
513f7b9504 Merge branch 'master' of github.com:RRZE-HPC/asmbench 2019-07-17 17:26:03 +02:00
Julian Hammer
a509ce3be0 fix for hypothetical uneven parallel sequence lengths 2019-07-17 17:25:14 +02:00
Julian Hammer
3f63e2c3bb version bump 2019-07-17 17:04:46 +02:00
Julian Hammer
7241bfc722 improved tp benchmarks by interleaving serial sequences 2019-07-17 17:02:41 +02:00
Julian Hammer
f422cbd1e6 passing frequency also to TP benchmark 2019-07-17 14:50:24 +02:00
Julian Hammer
80ae3a6036 Merge branch 'master' of github.com:RRZE-HPC/asmbench 2019-07-17 11:53:08 +02:00
Julian Hammer
242daeead3 increased parallel factor for load tp benchmark 2019-07-17 11:52:31 +02:00
Julian Hammer
79544033f0 graceful failing IACA analysis errors with -vvv 2019-07-16 13:22:08 +02:00
Julian Hammer
a89a96046f some frequency fixes 2019-07-16 10:20:24 +02:00
Julian Hammer
164eb03bf4 added (-f|--frequency) argument to cli of oldjit and asmbench 2019-07-16 09:38:42 +02:00
Julian Hammer
0e8157c594 fixed LD LAT result metric problem 2019-07-15 18:02:20 +02:00
Julian Hammer
7961659d85 adde console script asmbench 2019-02-01 13:49:18 +01:00
Julian Hammer
d218b55a50 better error message 2019-02-01 10:50:44 +01:00
Julian Hammer
f3ed5a0695 added sc18 src poster and summary 2019-01-17 13:41:20 +01:00
Julian Hammer
dd882088da removed irrelevant test files v0.1.2 2018-09-25 12:34:37 +02:00
Julian Hammer
54c485576f readded lost folder 2018-09-25 12:33:27 +02:00
Julian Hammer
5f3217c59f final commit for poster publication 2018-09-25 10:55:07 +02:00
Julian Hammer
21fb4d1c4d changed to rst readme 2018-09-25 10:37:53 +02:00
Julian Hammer
2cd25b4bcd added manifest file 2018-09-25 10:31:23 +02:00
Julian Hammer
dbbd37585a renamed to asmbench 2018-09-25 10:23:40 +02:00
Julian
3033246d4e Create sc18src_artifact_appendix.md 2018-09-17 16:04:33 +02:00
Julian Hammer
515b28cb4e added src extras 2018-08-02 10:49:49 +02:00
Julian Hammer
c5f82d8ec2 reordered instructions 2018-07-30 10:04:03 +02:00
Julian Hammer
0f6c440421 added div to benchmark 2018-07-30 10:02:07 +02:00
Julian Hammer
6172203aaf made repeat and elapstime configurable 2018-07-30 09:53:13 +02:00
Julian Hammer
cc51b232c2 workaround from ryzen / xen - llvm bug 2018-07-27 16:44:43 +02:00
Julian Hammer
d9bc86456d IACA report now available with -vvv 2018-07-27 14:40:16 +02:00
Julian Hammer
60ed2d2877 added zero runtime handling 2018-07-27 12:43:52 +02:00
Julian Hammer
e4c5b66741 handling overflow error in argument preperation 2018-07-27 12:30:18 +02:00
Julian Hammer
50c4da06f6 updated output 2018-07-27 11:00:22 +02:00
Julian Hammer
0dbf6b655b updated parameters 2018-07-26 16:35:32 +02:00
Julian Hammer
55ac8a142d added initial SRC script 2018-07-26 16:03:34 +02:00
Julian Hammer
6358dea1b0 added LICENSE and updated README.md 2018-07-24 13:12:53 +02:00
Julian Hammer
305f00f859 reduced register use in throughput mode 2018-07-11 16:23:22 +02:00
Julian Hammer
6786c399ad added support for iaca analysis 2018-07-10 14:26:19 +02:00
Julian Hammer
3dac7d6795 added cli support for serialized instructions 2018-07-05 14:25:15 +02:00