- 主题:Intel的 cripple AMD 手段
Intel MKL and other programs generated by the Intel C++ Compiler improve performance with a technique called function multi-versioning: a function is compiled or written for many of the x86 instruction set extensions, and at run-time a "master function" uses the CPUID instruction to select a version most appropriate for the current CPU. However, as long as the master function detects a non-Intel CPU, it almost always chooses the most basic (and slowest) function to use, regardless of what instruction sets the CPU claims to support. This has netted the system a nickname of "cripple AMD" routine since 2009.
--
FROM 222.128.190.*
哈哈，太损了
【 在 Heston 的大作中提到: 】
: Intel MKL and other programs generated by the Intel C++ Compiler improve performance with a technique called function multi-versioning: a function is compiled or written for many of the x86
--
FROM 219.143.131.*
为什么不用GCC编译器呢
【 在 Heston 的大作中提到: 】
: Intel MKL and other programs generated by the Intel C++ Compiler improve performance with a technique called function multi-versioning: a function is compiled or written for many of the x86 instruction set extensions, and at run-time a "master function" uses the CPUID instruction to select a version most appropriate for the current CPU. However, as long as the master function detects a non-Intel CPU, it almost always chooses the most basic (and slowest) function to use, regardless of what instruction sets the CPU claims to support. This has netted the system a nickname of "cripple AMD" routine since 2009.
--
FROM 221.234.149.*
ivf不知道会不会也这样
我用ivf编译的程序，跑mpi多线程在amd的机子还是碾压intel的
【 在 Heston 的大作中提到: 】
: Intel MKL and other programs generated by the Intel C++ Compiler improve performance with a technique called function multi-versioning: a function is compiled or written for many of the x86 instruction set extensions, and at run-time a "master function" u
: ses the CPUID instruction to select a version most appropriate for the current CPU. However, as long as the master function detects a non-Intel CPU, it almost always chooses the most basic (and slowest) function to use, regardless of what instruction sets
: the CPU claims to support. This has netted the system a nickname of "cripple AMD" routine since 2009.
: ...................
--
FROM 113.100.117.*
牛逼啊
【 在 Heston 的大作中提到: 】
Intel MKL and other programs generated by the Intel C++ Compiler improve performance with a technique called function multi-versioning: a function is compiled or written for many of the x86 instruction set extensions, and at run-time a "master function" uses the CPUID instruction to select a version most appropriate for the current CPU. However, as long as the master function detects a non-Intel CPU, it almost always chooses the most basic (and slowest) function to use, regardless of what instruction sets the CPU claims to support. This has netted the system a nickname of "cripple AMD" routine since 2009.
--
FROM 61.51.197.*
那些买线程撕裂者做工作站和epyc服务器的怎么办？
【 在 Heston 的大作中提到: 】
: Intel MKL and other programs generated by the Intel C++ Compiler improve performance with a technique called function multi-versioning: a function is compiled or written for many of the x86 instruction set extensions, and at run-time a "master function" uses the CPUID instruction to select a version most appropriate for the current CPU. However, as long as the master function detects a non-Intel CPU, it almost always chooses the most basic (and slowest) function to use, regardless of what instruction sets the CPU claims to support. This has netted the system a nickname of "cripple AMD" routine since 2009.
--
FROM 115.199.116.*
上面链接说了，mkl 2019以前的版本可以通过1个环境变量即可在epyc起飞。mkl 2022之后则稍折腾些，也能起飞
【 在 wwpty 的大作中提到: 】
: 那些买线程撕裂者做工作站和epyc服务器的怎么办？
--
FROM 117.128.4.*
MKL真的是臭名昭著。而且还和anaconda做了py交易后臭不要脸地绑定进官方包了。
--
FROM 180.164.55.*