Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

add fp8 defination #1258

Merged
merged 67 commits into from
Oct 29, 2024
Merged

add fp8 defination #1258

merged 67 commits into from
Oct 29, 2024

Conversation

uranus0515
Copy link
Contributor

add support for fp8 in k80

guodongliang and others added 26 commits October 8, 2024 15:19
rechange for performance in rvv
a13d346
@zhangyang2057
Avoid illegal instruction for both concat and slice with riscv gcc.
b0e7834
change back to 4100 for server12
7742941
@zhangyang2057
Merge branch 'dev/3.0' into feature/ntt_benchmark_roofline_4
bf944a7
@zhangyang2057
Modify timing from cycle to time of benchmark_ntt_matmul_primitive_size.
8f09a07
add fp8 defination
e323ad0
@uranus0515 @github-actions
Apply code-format changes
dad0377
add Quantizer match for Matmul
eb85032
Merge branch 'feature/fp8_e4m3_quant' of https://github.com/kendryte/…
00d7eb5 
…nncase into feature/fp8_e4m3_quant
@uranus0515 @github-actions
Apply code-format changes
08c6062
@zhangyang2057
Update matmul roofline for riscv64.
32c3d5c
add support for fp8
1ca77b7
Merge branch 'feature/fp8_e4m3_quant' of https://github.com/kendryte/…
d221c96 
…nncase into feature/fp8_e4m3_quant
@uranus0515 @github-actions
Apply code-format changes
4220a51
add support for function
19ea087
Merge branch 'feature/fp8_e4m3_quant' of https://github.com/kendryte/…
30ebff0 
…nncase into feature/fp8_e4m3_quant
add support for convert fp8 and half
96c597c
add type convert for cast
4470ae9
add support for fp8 matmul eval
a306418
@zhangyang2057
Add Cast.
4bfd8bc 
- support float/int32_t, flat/uint32_t, float/bool cast
- Add ukernel and rvv optimization
- Add ctest and benchmark test.
@zhangyang2057
uniform NTT_VL macro.
ca63b33
add support for evalation matmul
00d9819
@zhangyang2057
try to fix macos build.
9924c00
add evalate checker
fb189a2
@uranus0515 @github-actions
Apply code-format changes
40fcf6f
compitible for xpu and other target
5c86d27
@github-actions GitHub Actions
Copy link

github-actions bot commented Oct 15, 2024 •
edited
Loading

Test Results

5 113 tests  ±0   5 113 ✅ ±0   1h 20m 7s ⏱️ + 3m 31s
    5 suites ±0       0 💤 ±0 
    5 files   ±0       0 ❌ ±0 

Results for commit 36ef27d. ± Comparison against base commit 9ee9431.

♻️ This comment has been updated with latest results.

uranus0515 and others added 27 commits October 17, 2024 07:00
@uranus0515 @github-actions
Apply code-format changes
7229690
revise typo
3cd88b5
@uranus0515 @github-actions
Apply code-format changes
bd85689
add support for F8E5M2
8ac4ea8
@uranus0515 @github-actions
Apply code-format changes
ea0ef4f
@zhangyang2057
Limit VLEN <= 256 for matmul_primitive_size.
6ebec6d
add type support for int8
44b9ed9
add eval checker temporary
01867b5
Merge branch 'feature/ntt_benchmark_roofline_4' into feature/fp8_e4m3…
bbffcb1 
…_quant
change for compile error
e806ee1
Merge branch 'feature/ntt_benchmark_roofline_4' into feature/fp8_e4m3…
ec0c972 
…_quant
revise compiler error for arch
a9a7eb8
Merge branch 'feature/ntt_benchmark_roofline_4' into feature/fp8_e4m3…
5bb30bf 
…_quant
@xhuohai
pack eval support float8
75054df
@xhuohai
fix matmul pack type check
26dc30a
@xhuohai
add more vector types
58b6233
@xhuohai
fix tile solver
7426c2e
@xhuohai
fix matmul vector type infer
0ed8266
@xhuohai
fix float8.cs
4bfb1bd
@xhuohai @github-actions
Apply code-format changes
63982b4
add benchmark for gather
12f999f
fallback
0b1ed3a
@xhuohai
fix decompose softmax
d815d80
@xhuohai
scalar of where to tensor
88cf9ce
@xhuohai
ntt::mammul policy of mkn
c8095c5
@xhuohai
fix auto dist with tuple output
eed2ec3
@xhuohai
fix type infer of ScatterND
36ef27d
@uranus0515 uranus0515 merged commit 96a2dbc into dev/3.0 Oct 29, 2024
34 checks passed
@uranus0515 uranus0515 deleted the feature/fp8_e4m3_quant branch October 29, 2024 10:22
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@xhuohai xhuohai xhuohai approved these changes

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
3 participants
@uranus0515 @xhuohai @zhangyang2057