[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <17b2ed36-3075-4888-8057-0a471e5df209@gmail.com>
Date: Wed, 18 Sep 2024 21:48:09 +0000
From: Hanabishi <i.r.e.c.c.a.k.u.n+kernel.org@...il.com>
To: John <therealgraysky@...ton.me>
Cc: Borislav Petkov <bp@...en8.de>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
Thomas Gleixner <tglx@...utronix.de>, Ingo Molnar <mingo@...hat.com>,
Dave Hansen <dave.hansen@...ux.intel.com>, Unknown <x86@...nel.org>,
"H. Peter Anvin" <hpa@...or.com>
Subject: Re: [PATCH] x86: add more x86-64 micro-architecture levels
On 9/18/24 21:14, John wrote:
> I am not sure. Are some of the other things -march=-x86-64-v3 driving them?
Looking up a full table, v3 adds more than just AVX.
x86-64-v2:
CMPXCHG16B
LAHF-SAHF
POPCNT
SSE3
SSE4_1
SSE4_2
SSSE3
x86-64-v3:
AVX
AVX2
BMI1
BMI2
F16C
FMA
LZCNT
MOVBE
OSXSAVE
x86-64-v4:
AVX512F
AVX512BW
AVX512CD
AVX512DQ
AVX512VL
Maybe some other enabled instructions could issue some benefit.
v4 seems to be useless for us though.
> As to the code you referenced re: disabling the SIMD extensions. Do you know why that is in place?
Not really. There is a link above pointing to a bug report discussing GCC quirks. I am not an expert in that.
One day, out of curiosity, I tried to override it and build the kernel with '-mavx' (free performance, yay!).
Well, it didn't even start and crashed immediately.
I don't know if something has changed since then, but I guess there are reasons.
Powered by blists - more mailing lists