2a064cfcc5
added new dynamic shared mem test
...
Change-Id: Ic2a12fc9bc5b67e85f1e6c6008f10c7c66388377
2017-02-09 18:07:48 -06:00
6fd3daed30
fixed hipFunction memory management
...
Change-Id: I7ebb323419bcd220ebd6466a8eb38e7bfdb1520a
2017-02-09 17:22:55 -06:00
14442b3ef7
Update hip_hcc packaging script
...
Change-Id: I2dc96545c472942bcf2bc8a2e88735f3ba26d1e1
2017-02-09 20:28:28 +05:30
5754d641e0
Disable failing directed tests
...
- hipTestDeviceSymbol
- hipTestConstant
- hipTestMallocKernel
Change-Id: Ibfe9fc0b8a59882f1de64b42e18777a7bd56ee97
2017-02-09 14:48:22 +05:30
d14b239dd3
Fix arch specific hcc extra libs env var
...
Change-Id: I7429da2f1cb98750d6a9601e7e5bde844a098487
2017-02-09 14:38:39 +05:30
9a1989193a
added architecture specific macros
...
1. Added __HIP_ARCH_GFXNUM__
2. Usage, -D__HIP_ARCH_GFX803__=1
Change-Id: I68b3a85d62cfab3a45d2b7a70cb3518ab2565236
2017-02-08 19:45:32 -06:00
79978182f9
Add pseudo code example for hip_bugs.md
...
Change-Id: Ia2af8e6165faeb3fbb81428e20d4dc5b19b2fa9e
2017-02-08 16:20:01 -06:00
55d92833fb
fixed hipcc for new compiler flags
...
Change-Id: I49ec059be20ff26b7482c84d91ab7a43826c6a8d
2017-02-08 14:06:01 -06:00
27d2fc99ca
include arch specific ir on fallback path
...
Change-Id: Ib04996aae2c21eb73ef2a9f6305915e0caccd704
2017-02-08 12:19:06 -06:00
01b66dd998
Fixed Hawaii link issues
...
1. Split hip_ir.ll to hip_hc.ll and hip_hc_gfx803.ll
a. hip_hc.ll contains arch generic ir implementations
b. hip_hc_gfx803.ll contains gfx803 (fiji, polaris) specific ir
2. HIPCC can now parse --amdgpu-target=*.
a. Usage: hipcc --amdgpu-target=gfx803 --amdgpu-target=gfx701
b. TODO: Convert to --amdgpu-target=gfx803,gfx701
3. With LLC in HCC able to generate native f16 isa, removed inline half asm math ops
4. Fixed threadfence and threadfence_block to use functions in rocdl
Change-Id: Ic9a9e3e04139b0d75d2c2a263c030ca77adc1019
2017-02-08 12:04:05 -06:00
9ba480037a
[HIPIFY] Readme.md update.
2017-02-08 19:50:05 +03:00
841510481e
[HIPIFY] Readme.md update.
...
+ Launching example.
+ Delimiter "--" is strongly recommended.
+ "-x cuda" option is mentioned as well.
2017-02-08 19:43:32 +03:00
05c8aa1cf5
Fixed HIP for C runtime
...
1. fixed constant memory test
2. added runtime gcc/g++ test
3. added fix for runtime with gcc/g++
Change-Id: Ie14dad6633411b188bdeea044e140b6d5beabe16
2017-02-07 13:15:36 -06:00
3664e8784f
fixed vector types for c
...
Change-Id: I2330e976060f9a038929cd51be07ae2ee98e81ce
2017-02-07 11:09:54 -06:00
55df1b6ff2
Command scripts for latency measurements
...
Change-Id: I8c28765a09fb0358447367939de524b12699a317
2017-02-07 15:03:46 +05:30
939909e96c
Disable broken test: hipC
...
Change-Id: I061aa125bbdc9f14bc870266ab0735593c861903
2017-02-06 13:00:50 +05:30
5e3d63c0a3
changed __global__ attribute
...
1. Moved around tests and added them to HIT
Change-Id: I5d75280c42a5af852670ebabc7305ee56721ec7b
2017-02-03 10:53:36 -06:00
2790e9a448
fixed symbol memcpy issue
...
Change-Id: I89d7401be51d194bcbf771020ba66e3d3b6a18f8
2017-02-01 17:54:59 -06:00
60013396f6
removed host math functions from math_functions.h
...
Change-Id: I90d8784e2d6b58c6fade9f0fa12c0db3ee417d3e
2017-01-27 17:38:43 -06:00
66dc2d42db
changed device code tests to work not to work as one
...
Change-Id: I0eec1eab19dda3b703bc3a0d778a6bbb2802a412
2017-01-27 09:20:14 -06:00
b37422d30b
fixed arch compiler flag
...
Change-Id: I19f7a0ea513d6b8059f0c10cd0e7a5ead510e246
2017-01-27 08:51:48 -06:00
f7ff199daa
fixed threadfence ir
...
Change-Id: Ia3afb54bdb50864e678d849608d72a3c321edba1
2017-01-27 08:42:26 -06:00
6ba848a40f
fixed kernel only compilation for module api
...
Change-Id: I567992fa9b87125318edba41fd82d2f7bc5504a1
2017-01-27 08:41:42 -06:00
3e21d55c98
hipMallocArray fixes for NV path
...
Change-Id: I1ca43e6bc0cd405998888005c20dfb1ea57003d5
2017-01-27 14:32:08 +05:30
2e1a66103a
more fix on hipmallocarray on NV path
...
Change-Id: I890a36cab10c101f4a112bc4567f765b318d486c
2017-01-26 12:30:52 -06:00
8404215658
fix missing semicolon on NV path
...
Change-Id: I3cfecb7bd534578a1f5a07ca9397092dcf01db07
2017-01-26 11:28:15 -06:00
0409bf639c
Add HIP_FAIL_SOC.
...
Fail sub-optimal-copies rather than perform them slowly.
SOC occur on async copy of unpinned memory, or P2P copy between GPUs
that are not peers.
2017-01-25 21:53:17 -06:00
1635b8f43f
Read HCC_OPT_FLUSH and optimize dispatch accordingly.
...
If HCC is in this mode, we can use less aggressive flushes in some
cases.
2017-01-25 21:50:52 -06:00
d75d0596bf
added device functions header to hip_runtime.h
...
Change-Id: I635931e1fbe4b7f0f64f3a126c0f1badcd6d234b
2017-01-25 08:14:30 -06:00
95677edabb
Add more hip_bug.md entry, regarding hang after hipLaunchKernel
...
Change-Id: I5800cb627179ec0e913cd36d332fb8c2994ab71e
2017-01-24 22:43:25 -06:00
f3da91de4e
Initial commit on hip_bugs markdown doc
...
Change-Id: I5a6915337b8664cfed9eaee9443c6e4406348574
2017-01-24 22:30:36 -06:00
9dff0782a4
Add debug tip to docs
2017-01-23 22:34:41 -06:00
813c189b33
Show dynamic shared mem usage not static.
2017-01-23 22:34:41 -06:00
0dabdeb01f
Move core env var processing to env.cpp
2017-01-23 22:34:41 -06:00
96eac67929
Add debug tips to docs
2017-01-23 22:34:41 -06:00
85d41dcd05
Add debug tips to docs
2017-01-23 22:34:41 -06:00
4586091dfe
Log error with ihipLogError. Cleans up CXL trace display.
2017-01-23 22:34:41 -06:00
138cf36547
Add HIP_IGNORE_HCC_VERSION.
...
Ignores strict checking of HCC and HIP version.
Can be useful when developing new HCC code.
2017-01-23 22:34:41 -06:00
4e3afa6514
added ir code sad u8
...
Change-Id: Ie0d454b3bb9a6c9a028c091ad3aa969719b02cc9
2017-01-20 17:21:51 -06:00
8ddec0426b
added driver_types.h and texture_types.h header files to hip
...
Change-Id: Ic3b2403f07d6767dadf83d6c278fd14e87f6acdb
2017-01-20 17:09:52 -06:00
22acd654cf
fixed hipArray issues
...
1. Fixed build issues produced from previous commit
2. Create new header files to manage data structures better
Change-Id: I704d82c196c1858ed7617d76e40612eb507d2aa0
2017-01-20 16:54:48 -06:00
7765469987
changes device functions documentation according to the supported apis
...
Change-Id: I47ac6bbde11d54d8265e0d27ec8cd5da4d03eb8e
2017-01-20 14:19:09 -06:00
b1eca6c855
added nvcc backend for hipArrays
...
1. Added hip_texture.h to hip_runtime_api.h as cuda does declare array runtime apis inside cuda_runtime_api.h
2. Added nvcc backend for hipArray runtime apis
3. Didn't test on nvidia platform (should work)
Change-Id: I1a14aef41840e4f55e5535132e3443a918b55967
2017-01-20 14:11:45 -06:00
b134a1a640
added more test coverage for vector data types
...
Change-Id: I9f57a8b597bd2ee4b265eadfd0859531497a6ada
2017-01-20 13:52:02 -06:00
f537d96633
fixed compilation issues for vector types and math functions
...
1. Added math_functions.h to hip_runtime.h
2. Changed operator overloading classifier static to static inline
3. Added vector types test for gpu
4. Seperated __host__ and __device__ for math functions in headers
Change-Id: I499862fad5d7b10da686da9011d7ecefe523f8e2
2017-01-20 09:49:11 -06:00
927ac3d81c
Add HIP_SYNC_HOST_ALLOC, HipReadEnv
2017-01-19 23:55:24 -06:00
8209320ef0
Change ihipDeviceSetState,ihipDevice* so it doesn't log error
...
Cleans up debug trace.
2017-01-19 23:55:24 -06:00
6ca2b289a2
added operator overloading for complex data types
...
Change-Id: Id96d5d000651914169f04497af6ff78ad96d846a
2017-01-19 15:15:25 -06:00
1f5d16afe7
Doc update - describe debug techniques
...
Also tweak sample to remove unneeded HIP_KERNEL_NAME.
Comment update
2017-01-19 12:40:45 -06:00
1c73e44ebe
Fix debug display for Module launch kernels
2017-01-19 12:40:45 -06:00
Aditya Atluri
Maneesh Gupta
pensun
Evgeny Mankov
Rahul Garg
Ben Sander