71eca27ffa
removing LD_PRELOAD for hip trace as the HIP library is now loading statically
...
Change-Id: I032c73feb9e8426f5b2cd5bd930e7160c3b07294
2019-12-13 01:55:34 -06:00
31790cc69d
LD_PRELOAD interference fixed
...
Change-Id: Ib44aa28a71326900f5a71bbb37dbd8049cc3e3a2
2019-12-12 11:41:00 -06:00
b0d56ba0eb
tracing on start option
...
Change-Id: Ie7a85da4fd865fbabd6b38ac1d12977ff2684114
2019-12-11 14:41:51 -06:00
28e59eb0da
hcc library loading workaround
...
Change-Id: I8b383f901235cfdfaba612a4effa4b9fcf12104d
2019-12-10 06:40:09 -06:00
389cb50854
trace parameters fix
...
Change-Id: I5cd8c897b9982f9abc9e36c18a4a488bd65fbd30
2019-12-10 04:33:29 -06:00
b74ae2f62d
kfd trace rocprof integration
...
Change-Id: Ie1f1679f50a973d071b81f3cf9fca570da9f12c3
2019-12-03 09:47:58 -06:00
c8eb0641bf
sgpr/vgpr reported sizes in terms of registers; lds size in bytes;
...
Change-Id: Id527687cc075a4e85ea4d0063ed8f336e34d2454
2019-11-27 15:19:14 -06:00
cafeee3f9d
fix RPM packaging, setting rpm_post as RPM_POST_INSTALL
...
Change-Id: I91790a3f2a8368940eba29cf1ff76728fc9ac3a1
2019-11-25 16:29:15 -06:00
f14753c195
results_mcopy.csv generating
...
Change-Id: I8774be4ca4ad23a11fc0ba3489d4da0cce0a77c8
2019-11-19 20:26:50 -06:00
60043d198b
code object tracking and v3 code object support
...
Change-Id: I081ada13f6364ea4401a97a485dedfa9bf8a45fe
2019-11-19 20:18:09 -06:00
a010540901
packet signal tracking fix, added 1 as a default
...
Change-Id: I50f17e90e0f4ff8584ba427fd1122d91aeb49fa0
2019-11-19 20:01:05 -06:00
dc05345a78
setting of granulated lds, sgpr, vgpr reserved by HW for the kernel
...
Change-Id: Ib3e60c0b97787f56f769136480a46a8560e00cc8
2019-11-11 18:11:16 -05:00
b9989fb33c
profiled cmd line quoting option
...
Change-Id: I576065eb289b59829978c7a2acf74799b68bc213
2019-11-11 18:10:57 -05:00
d8560ae0b3
enable rocprofiler kernels trace; fixing signal value for proxy signal; fixing scope options in input txt format;
...
Change-Id: Ife63c4091e565a2158e292c7acf9718085c709be
2019-11-01 20:44:02 -05:00
00ca6d8a78
changing trace control to <delay:length:rate>
...
Change-Id: I6eca07257ab8d06972e4eba6ee04ae1200b50fe9
2019-10-10 19:04:30 -05:00
8ebf2fc5fa
roctx rocprof integration
...
Change-Id: Ia5e3a60c8d89ef0b04a83f09be361454a95e9689
2019-10-09 23:39:59 -05:00
484a8402e1
custom metrics file rocprof option
...
Change-Id: I3914740a7640a291558cbe70ca1bdb0400ff7cee
2019-10-09 18:03:04 -05:00
c783e62230
--trace-period option
...
Change-Id: I29bce7763e62659d22d6abe144b8d970ab4f41c3
2019-10-09 18:34:48 -04:00
d8bd1f5ae5
fixed querying of available metrics to use correct agnet name
...
Change-Id: Ic2c82d5cba760ba5b0caa373e5309f17c908b8e9
2019-10-03 13:30:33 -04:00
363107b09e
Merge "tests: replace 'bc' with 'awk' to be more portable" into amd-master
2019-09-17 17:39:21 -04:00
d2bb3be7f9
fixing build instructions
...
Change-Id: I7a8e2a04f7a048864f4f8fd89a92fceccfcc867f
2019-09-13 16:41:15 -05:00
6f6208f06d
scripts fixes
...
Change-Id: I1d51bf4fe2ba62f26bafc251ae702e1064a3e5e3
2019-09-11 19:22:26 -05:00
51bdb2ce02
tests: replace 'bc' with 'awk' to be more portable
...
Calbulator package 'bc' is not always installed, occasionally
breaking the test scripts. As a fix, this patch gets rid of
'bc' and uses 'awk' instead.
Change-Id: I5f4dbf018d97581a626cf95d147c415755477771
2019-08-26 10:53:52 -04:00
e5b938b5fb
unordered metrics expressions config
...
Change-Id: I998423046abfd7af45f28071911276cce1c4f3da
2019-08-25 05:08:55 -05:00
6fa4f1dc66
Merge "tests: improve scripts on rocm v2.6, and do more check" into amd-master
2019-08-23 15:02:06 -04:00
68bf049d49
tests: improve scripts on rocm v2.6, and do more check
...
'rocminfo' output format on v2.6 is slightly different, and thus
cache capacity and cache linesize can be incorrectly obtained.
Example: 'L1: 16KB' vs. 'L1: 16(0x10) KB'. This patch fixes the
formatting issue, and further does more check on profiled .csv
file.
Change-Id: I934ee3613a79fa8acc431a394e88a9e09833311b
2019-08-23 12:49:59 -04:00
4f4e02de7a
Merge "build: slightly change files to pass compilation" into amd-master
2019-08-23 11:44:44 -04:00
aecc7abc76
Mem32Bwrites -> MemWrites32B renaming
...
Change-Id: I86f570b97178caa3d3d5fcbc5ecc716216dad504
2019-08-22 22:34:36 -05:00
dff3256dd5
mi100 metrics fix
...
Change-Id: I862578a2b7827a88cc6926d636d5eb9a8320a244
2019-08-22 22:30:48 -05:00
ee917344ff
build: slightly change files to pass compilation
...
Compiled using Clang and encountered errors on ambiguous pointer
type, implicit conversion NULL to bool, and unused function, etc.
As a fix, this patch slightly updates the related codes.
Change-Id: I4a73d66293313dda36749036f5c73aaf2a445231
2019-08-22 23:29:16 -04:00
4a9c36b32c
multiple processes tracing support
...
Change-Id: Ic9afcdfdf9033f35bdfb78007ebf96eec5f0de44
2019-08-21 11:53:27 -05:00
72fdf1afcb
readme update
...
Change-Id: Ic09aa90e9b6d584aa1dca4e54f6c8629493aa501
2019-08-10 00:15:57 -05:00
c635e3a25e
inheritance fix
...
Change-Id: I1ff0539b658c5e742a05085b8b5514ffd2c10f47
2019-08-08 18:00:51 -05:00
82ecf9e7a4
overwriting on inheritance
...
Change-Id: I6faf79874fc14e898c75972aa2f372e228105f7b
2019-08-08 01:54:29 -05:00
09eed4f35b
gfx908 events update
...
Change-Id: Ieb7c72f66b7e5e5dfc29d97650a29722a6276a31
2019-08-07 19:44:31 -05:00
855a801ae7
adding gfx908 namespace
...
Change-Id: Ifd45385777be5066c8bfef23ee24bbabadf4dffa
2019-07-27 03:06:10 -05:00
9678e5f680
fixing libraries pathes
...
Change-Id: Ife3186ed10309ac674dd8f112b3a6647dce06c7b
2019-07-26 22:13:58 -05:00
cb1a7f6481
Merge "sys trace support" into amd-master
2019-07-24 16:46:34 -04:00
8092bd9b2b
Merge "tests: make tests print msg and exit on errors" into amd-master
2019-07-24 14:20:25 -04:00
ed68e4c432
Merge "tests: remove TCC_EA_* counters to support gfx8" into amd-master
2019-07-24 14:20:14 -04:00
86c6f93536
sys trace support
...
Change-Id: I7297a0bf73c2895b1751c97f9ee9bbd1c1c2aa93
2019-07-24 13:15:25 -05:00
87e46dde4e
tests: make tests print msg and exit on errors
...
Change-Id: Ie5af942551a6970a7414c30988a47ed57a9baf27
2019-07-23 17:18:54 -04:00
710a9de411
Merge "integrating memory_validation bench in cmake build" into amd-master
2019-07-23 14:49:50 -04:00
657241e848
Merge "xml: add extra events for gfx906 GPUs" into amd-master
2019-07-23 12:40:39 -04:00
6c5ca62575
integrating memory_validation bench in cmake build
...
Change-Id: I90ad85c7b1517c162804a00d5abb75007bf831fe
2019-07-23 11:14:44 -05:00
6152563612
tests: remove TCC_EA_* counters to support gfx8
...
TCC_EA_* counters are only available on gfx9 devices, and thus
the memory validation tests fail on gfx8. To fix the issue,
this patch removes TCC_EA_* metrics and keeps only the ones
supported by both gfx8 and gfx9.
Change-Id: Ib3889435b88f2e6970981304a72cee75d5343d4b
2019-07-23 05:00:28 -05:00
1d9b6b4d4f
Merge "tests: add top-level Makefile and run.sh" into amd-master
2019-07-22 16:09:47 -04:00
a5a56425bb
tests: add top-level Makefile and run.sh
...
Change-Id: Ieaef51422c11ec27d56ff19953f9a5dd1f563353
2019-07-18 05:21:20 -05:00
4ef7d7d156
sys trace adding
...
Change-Id: I8d0a3a34e14f897e635a42faa42eedc1a1f86ace
2019-07-18 00:41:49 -05:00
ca9a714b77
xml: add extra events for gfx906 GPUs
...
Compard to gfx900, e.g., Vega 10, gfx906 adds extra counter events.
A typical difference is on TCC-EA that gfx906 (e.g., Vega 20) has 2
EAs per TCC, while only one single EA/TCC on gfx900. As such,
additional counters must be profiled to get correct results. This
patch adds one extra events to specifically handle gfx906.
Change-Id: Id6c9d37548a102c80bbfddcfa11e77d20f17431a
2019-07-16 12:56:45 -05:00
Evgeny
Xianwei Zhang