kvcache-ai/ktransformers

Fork 0

mirror of https://github.com/kvcache-ai/ktransformers.git synced 2026-11-05 01:07:28 +00:00

[GH-ISSUE #520] CPU不支持AMX指令，哪个版本能用？ #372

New issue

Closed

opened 2026-10-29 14:38:43 +00:00 by bsakaguchi · 6 comments

bsakaguchi commented

2026-10-29 14:38:43 +00:00

Owner

Originally created by @cowcomic on GitHub (Feb 20, 2025).
Original GitHub issue: https://github.com/kvcache-ai/ktransformers/issues/520

目前使用的0.2.1的docker启动的
报如下错

loading blk.0.attn_q_a_norm.weight to cuda:0
loading blk.0.attn_kv_a_norm.weight to cuda:0
loading blk.0.attn_kv_b.weight to cuda:0
loading blk.0.attn_norm.weight to cuda:0
loading blk.0.ffn_norm.weight to cuda:0
loading blk.1.attn_q_a_norm.weight to cuda:0
loading blk.1.attn_kv_a_norm.weight to cuda:0
loading blk.1.attn_kv_b.weight to cuda:0
loading blk.1.attn_norm.weight to cuda:0
loading blk.1.ffn_norm.weight to cuda:0
loading blk.2.attn_q_a_norm.weight to cuda:0
loading blk.2.attn_kv_a_norm.weight to cuda:0
loading blk.2.attn_kv_b.weight to cuda:0
loading blk.2.attn_norm.weight to cuda:0
loading blk.2.ffn_norm.weight to cuda:0
loading blk.3.attn_q_a_norm.weight to cuda:0
loading blk.3.attn_kv_a_norm.weight to cuda:0
loading blk.3.attn_kv_b.weight to cuda:0
Illegal instruction (core dumped)

issue中查到是amx指令集导致的
我看 320 这个issue提供了个解决方案，但好像只针对CPU支持AMX指令集的生效
我试了没用
直接报
“Fail to do XFEATURE_XTILEDATA”

求助：不支持AMX指令集的CPU用哪个版本能运行起来？

Originally created by @cowcomic on GitHub (Feb 20, 2025). Original GitHub issue: https://github.com/kvcache-ai/ktransformers/issues/520 目前使用的0.2.1的docker启动的报如下错 ``` loading blk.0.attn_q_a_norm.weight to cuda:0 loading blk.0.attn_kv_a_norm.weight to cuda:0 loading blk.0.attn_kv_b.weight to cuda:0 loading blk.0.attn_norm.weight to cuda:0 loading blk.0.ffn_norm.weight to cuda:0 loading blk.1.attn_q_a_norm.weight to cuda:0 loading blk.1.attn_kv_a_norm.weight to cuda:0 loading blk.1.attn_kv_b.weight to cuda:0 loading blk.1.attn_norm.weight to cuda:0 loading blk.1.ffn_norm.weight to cuda:0 loading blk.2.attn_q_a_norm.weight to cuda:0 loading blk.2.attn_kv_a_norm.weight to cuda:0 loading blk.2.attn_kv_b.weight to cuda:0 loading blk.2.attn_norm.weight to cuda:0 loading blk.2.ffn_norm.weight to cuda:0 loading blk.3.attn_q_a_norm.weight to cuda:0 loading blk.3.attn_kv_a_norm.weight to cuda:0 loading blk.3.attn_kv_b.weight to cuda:0 Illegal instruction (core dumped) ``` issue中查到是amx指令集导致的我看 [320](https://github.com/kvcache-ai/ktransformers/issues/320) 这个issue提供了个解决方案，但好像只针对CPU支持AMX指令集的生效我试了没用直接报 “Fail to do XFEATURE_XTILEDATA” 求助：不支持AMX指令集的CPU用哪个版本能运行起来？

bsakaguchi

2026-10-29 14:38:43 +00:00

closed this issue
added the
stale
label

bsakaguchi commented

2026-10-29 14:38:43 +00:00

Author

Owner

@yileld commented on GitHub (Feb 20, 2025):

官方的release页里有其他指令集的wheel包

@yileld commented on GitHub (Feb 20, 2025): 官方的release页里有其他指令集的wheel包

bsakaguchi commented

2026-10-29 14:38:44 +00:00

Author

Owner

@ipc-robot commented on GitHub (Feb 20, 2025):

请问这个项目现在是需要支持AMX指令集的机器才能最佳运行吗？

@ipc-robot commented on GitHub (Feb 20, 2025): 请问这个项目现在是需要支持AMX指令集的机器才能最佳运行吗？

bsakaguchi commented

2026-10-29 14:38:44 +00:00

Author

Owner

@Azure-Tang commented on GitHub (Feb 20, 2025):

除了0.3 preview都能跑。用docker的话在里面重新pull一下最新代码编译一下，我们打包的机器有amx所以把amx也编进去了

@Azure-Tang commented on GitHub (Feb 20, 2025): 除了0.3 preview都能跑。用docker的话在里面重新pull一下最新代码编译一下，我们打包的机器有amx所以把amx也编进去了

bsakaguchi commented

2026-10-29 14:38:44 +00:00

Author

Owner

@codedoves commented on GitHub (Feb 21, 2025):

支持avx512么？重新安装一下ktranfoemer,换成对应版本avx512的试试

@codedoves commented on GitHub (Feb 21, 2025): 支持avx512么？重新安装一下ktranfoemer,换成对应版本avx512的试试 ![Image](https://github.com/user-attachments/assets/fdbebc2d-eee1-4220-bc87-e3c0bf20f385)

bsakaguchi commented

2026-10-29 14:38:44 +00:00

Author

Owner

@cowcomic commented on GitHub (Feb 23, 2025):

感谢，我试试

@cowcomic commented on GitHub (Feb 23, 2025): 感谢，我试试

bsakaguchi commented

2026-10-29 14:38:44 +00:00

Author

Owner

@ErvinXie commented on GitHub (Dec 11, 2025):

This issue has been automatically closed due to inactivity for 60 days.

If you believe this issue is still relevant, please feel free to reopen it with additional information or context.

Thank you for your contribution to KTransformers!

此 issue 因超过 60 天无活动而被自动关闭。

如果您认为此问题仍然存在，请随时重新打开并提供更多信息或上下文。

感谢您对 KTransformers 的贡献！

@ErvinXie commented on GitHub (Dec 11, 2025): This issue has been automatically closed due to inactivity for 60 days. If you believe this issue is still relevant, please feel free to reopen it with additional information or context. Thank you for your contribution to KTransformers! --- 此 issue 因超过 60 天无活动而被自动关闭。如果您认为此问题仍然存在，请随时重新打开并提供更多信息或上下文。感谢您对 KTransformers 的贡献！