GithubStar/ncnn
1
0
Fork 0
mirror of https://github.com/Tencent/ncnn.git synced 2026-10-18 13:42:18 +08:00
Code Issues Packages Projects Releases Wiki Activity
3,968 Commits 4 Branches 49 Tags
master
Commit Graph

236 Commits

Author SHA1 Message Date
nihui 60bfa40044 sdpa flash attention x86 optimization, llm models enable gqa, dispatch for avx2 bfloat2float (#6923) 
* enable gqa
* update perf test
 2026-09-01 17:35:42 +08:00
nihui 7f162381ab kvcache allocator (#6901) 
* decode benchmark
* reject kvcache batch
* enable mha vulkan gemm out_elempack 1
 2026-08-17 15:44:30 +08:00
nihui 4733fe3173 vktransfer command may reset and disable fp16 in llm benchmarks (#6887) 2026-08-04 17:18:10 +08:00
nihui 92b73d6665 x86/arm/mips/loongarch/riscv gemm block quantization optimization (#6831) 
* benchncnn_llm int8 models
 2026-08-04 14:34:22 +08:00
nihui f6f734f44d vulkan int8 kernels (#6751) 
* vulkan: add int8 kernels for core layers
Implement Vulkan int8 paths for convolution, convolutiondepthwise, innerproduct, and gemm with scalar fallback and integer dot product shader branches. Add goal.md and enable existing int8 unit tests for Vulkan coverage.
* enable int8 gpu benchmark
 2026-06-17 17:39:49 +08:00
nihui 619ea8e0b2 change llm benchmark seqlen from 1024 to 256 (#6738) 2026-05-22 17:04:32 +08:00
nihui 244572aeb8 fix simplemath build, trigger ci on benchmark changes (#6722) 2026-05-14 11:27:58 +08:00
nihui 13bca30c00 benchncnn_llm (#6711) 2026-05-13 11:38:29 +08:00
nihui 155ab53695 move model files into sub directory (#6710) 2026-05-09 16:32:02 +08:00
nihui d4f599f2cd benchmark: use Gemm for vision transformer linear layers (#6709) 2026-05-09 14:29:54 +08:00
Kenji Mouri / 毛利 研二 f4e9a78291 Add benchmark results for several instances from Microsoft Azure. (#6552) 2026-03-30 10:21:46 +08:00
Ratizux c8e774f247 add Qualcomm Snapdragon X Elite benchmark (#6535) 2026-02-07 21:56:36 +08:00
Apeiria 6b4f6f95e5 [feat](benchmark): replace load_param with load_param_mem (#6286) 
* [feat](benchmark): replace load_param with load_param_mem

* [docs](benchmark): update benchmark docs

* refactor(benchmark): remove model registry and optimize param loading

* fix(benchmark): correct README param description

* refactor(benchmark): rename header file

* update doc

---------

Co-authored-by: nihui <nihuini@tencent.com>
Co-authored-by: nihui <shuizhuyuanluo@126.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
 2026-01-08 14:23:00 +08:00
SimoSbara d0afafe7bd Update README with Arduino UNO Q benchmark results (#6454) 2025-12-15 10:14:10 +08:00
張小凡 b8a1c92cce fix format specifier for size_t to support 64-bit values (#6315) 
---------

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
 2025-09-12 14:00:27 +08:00
nihui b5bc05e2e8 drop vulkan shader pack8 (#6270) 2025-08-22 11:05:59 +08:00
WANG KE 4d8220ce22 Skip int8 model on GPU and merge from upstream (#6174) 2025-07-12 17:06:46 +08:00
nihui 171b9d1bba use spdx license header, copyright Tencent (#6152) 2025-07-08 11:01:29 +08:00
nihui 24a3b99f1f drop layer support_image_storage and option use_image_storage (#6126) 
* fix pyncnn build
 2025-06-16 11:23:41 +08:00
Yingxi 46b3de348c Add MUSE Pi Pro Spacemit M1 benchmark. (#6074) 2025-05-26 22:35:55 +08:00
青菜萝 卜冬瓜 edfb2c6bee Add MacBook Air (13-inch, M3, 2024) benchmark. (#5645) 2025-04-18 10:27:11 +08:00
TheSnowfield 55e5c0ba05 benchncnn: add radxa orion o6 benchmark (#5943) 2025-03-25 07:56:53 +08:00
GIBEREZ cdcbb3b836 benchncnn: add vim3 vulkan benchmark (#5821) 2024-12-14 17:34:00 +08:00
Kelun Lei 07196eee2e benchmark: add Kunpeng 920 7260 (#5606) 2024-08-16 10:50:26 +08:00
UOPiceman 74037b49f8 Add Axera AX630C benchmark (#5559) 2024-07-08 12:59:13 +08:00
inspireMeNow 1225793e56 benchmark: add Snapdragon 765G and CVITEK SG2000 (#5555) 2024-07-08 11:07:11 +08:00
青菜萝 卜冬瓜 9f62738e7c Add RaspberryPi 5 CPU Overclock benchmark. (#5561) 2024-07-08 11:01:12 +08:00
py1066 317635b2fb add OrangePicm4 (#5551) 2024-07-02 12:37:53 +08:00
inspireMeNow e48c8aa21c benchmark: add OrangePi 5 Plus (#5550) 2024-07-01 10:46:18 +08:00
青菜萝 卜冬瓜 5d4f0213ce Add Snapdragon 888 benchmark. (#5548) 2024-07-01 10:44:58 +08:00
sakria9 c8e6e18918 benchmark: add EPYC 7742 and V100 (#5528) 2024-06-23 23:53:16 +08:00
CharlieYu b786af56f8 benchmark: add raspberry pi 5 benchmark after GPU overclock (#5518) 2024-06-19 15:52:22 +08:00
nihui e389a15846 update qcom855plus benchmark 2024-06-12 15:08:15 +08:00
Q-engineering 5fdbb54b7f Add Radxa Zero 3W to benchmark (#5321) 2024-01-30 14:15:33 +08:00
Darren Cheng 10fd242741 add remipi benchmark (#5319) 2024-01-29 17:01:40 +08:00
HalfSweet 66b26b6ce8 add PhytiumPi ncnn benchmark (#5312) 2024-01-26 13:08:27 +08:00
Molly Sophia 09f15e6980 Add Dimensity 9300 MT6989 ncnn benchmark (#5284) 
Signed-off-by: Molly Sophia <mollysophia379@gmail.com>
 2024-01-15 10:45:19 +08:00
wonderfullook 28078b7ad6 add orangepi zero2 ncnn benchmark (#5277) 2024-01-12 15:30:59 +08:00
mizu-bai 2ac1e776dd Add Xeon Phi 3120A results (#5276) 2024-01-12 15:30:36 +08:00
mizu-bai 237f45ff3e Update AWS and OneCloud benchmark results with ncnn 20240102 (#5257) 
* Update OneCloud results with 20240102

* Update AWS c5.4xlarge Instance results with ncnn 20240102

* Add compiler info for OneCloud
 2024-01-08 11:01:42 +08:00
nihui ded0b78bb2 fix nvidia vulkan crash on exit (#5234) 2023-12-20 11:32:53 +08:00
wzyforgit 9023ecf52c update benchmark for visionfive2 (#5217) 2023-12-15 15:13:08 +08:00
Q-engineering 070dac9955 Benchmark results Jetson Orin Nano, RPi 5 (refreshed), and a simple ranking tool (#5216) 2023-12-12 11:15:17 +08:00
Han Gao/Revy/Rabenda 3fcb8a7263 Add 2K2000 benchmark (#5214) 2023-12-11 11:45:54 +08:00
FantasyGmm 3cff53c30a add Jetson TX2 benchmark (#5213) 2023-12-11 10:39:29 +08:00
mahiru 5c432d5985 add the benchmark of 8gen2 (#5180) 2023-11-28 10:29:59 +08:00
FantasyGmm a029587725 Update RPi 5 GPU Benchmark Result (#5160) 
use mesa-23.3-rc4 driver
 2023-11-21 14:15:57 +08:00
FantasyGmm 22c990ac65 Add Raspberry Pi 5 GPU Benchmark (#5152) 2023-11-15 17:09:04 +08:00
Justin Fung 465debe9bb Add print statements for 4 dimensions benchmark (#5148) 2023-11-15 11:43:59 +08:00
mizu-bai c96bc08307 Add OneCloud Amlogic S805 (#5120) 2023-10-31 17:34:34 +08:00