[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <KcA5fQwm2CimaycjqPqz-HP7y7Dyx3MbSNUc2F6eYqN5T48hLrFAUwiajYvOaat8Apn-dZvzQ2RAp2Ln-9BE2s1uYn7LwFdTX7NLqbBkC5k=@proton.me>
Date: Wed, 18 Sep 2024 21:14:24 +0000
From: John <therealgraysky@...ton.me>
To: Hanabishi <i.r.e.c.c.a.k.u.n+kernel.org@...il.com>
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 Wednesday, September 18th, 2024 at 4:25 PM, Hanabishi <i.r.e.c.c.a.k.u.n+kernel.org@...il.com> wrote:
> Even better then! Could you please explain where the performance gains should come from, considering that the kernel force disables all SIMD extensions?
> https://github.com/torvalds/linux/blob/4a39ac5b7d62679c07a3e3d12b0f6982377d8a7d/arch/x86/Makefile#L67-L80
>
> I.e. if we won't have them anyway, what gives?
I am not sure. Are some of the other things -march=-x86-64-v3 driving them? I will say that these timed benchmarks have been consistently reproducible for me. My code for the benchmark script is in that github repo as well if you would like to give it a whirl.
As to the code you referenced re: disabling the SIMD extensions. Do you know why that is in place?
Powered by blists - more mailing lists