From nobody Wed Sep 13 04:26:29 2023 X-Original-To: freebsd-fs@mlmmj.nyi.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2610:1c1:1:606c::19:1]) by mlmmj.nyi.freebsd.org (Postfix) with ESMTP id 4RlnTc5kGcz4tfNy for ; Wed, 13 Sep 2023 04:26:44 +0000 (UTC) (envelope-from rincebrain@gmail.com) Received: from mail-wm1-x32b.google.com (mail-wm1-x32b.google.com [IPv6:2a00:1450:4864:20::32b]) (using TLSv1.3 with cipher TLS_AES_128_GCM_SHA256 (128/128 bits) key-exchange X25519 server-signature RSA-PSS (4096 bits) server-digest SHA256 client-signature RSA-PSS (2048 bits) client-digest SHA256) (Client CN "smtp.gmail.com", Issuer "GTS CA 1D4" (verified OK)) by mx1.freebsd.org (Postfix) with ESMTPS id 4RlnTc45kCz3PlG for ; Wed, 13 Sep 2023 04:26:44 +0000 (UTC) (envelope-from rincebrain@gmail.com) Authentication-Results: mx1.freebsd.org; none Received: by mail-wm1-x32b.google.com with SMTP id 5b1f17b1804b1-401da71b83cso71838415e9.2 for ; Tue, 12 Sep 2023 21:26:44 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20221208; t=1694579202; x=1695184002; darn=freebsd.org; h=cc:to:subject:message-id:date:from:in-reply-to:references :mime-version:from:to:cc:subject:date:message-id:reply-to; bh=0cXHK2561DoEf0VSILeDrovlt4Qq+86OLi+DqmuGXJ0=; b=svIFlbT2tOC2Bt1REJcMcMSYAj7pfxeL9cn2Eg9b5P7g2rxHp7IhD/gRmhT/JiJ3Jo eWH3rACVuPawmY1XuEIhIWAGAKwT6/s47QIk8qkvTIt3hHnAxtpNAkyaLLtZUWl+U/zC abENNsT2ZrFzFqLrJkMUmhxYxOfUfTALsCiwvYz2pDPoZH7kbmhryg1tp0DRTbH3eksb 9YBq6r760PQxzTyUcB4RFIjZdxWBD2Dxk+IPw1MV3j2zXQFRqMBvJOZaqyyCjG9ITVyV f4o66+EU7Iq02xkR1n6BeHqqxDMZK8Wc92jFhPhDZU3Vb3BeoLiJpnlLFM1/v1gRe7LT iz0Q== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1694579202; x=1695184002; h=cc:to:subject:message-id:date:from:in-reply-to:references :mime-version:x-gm-message-state:from:to:cc:subject:date:message-id :reply-to; bh=0cXHK2561DoEf0VSILeDrovlt4Qq+86OLi+DqmuGXJ0=; b=CVWWxuORKs0x8Lrx7LdmWzTh18ACBXiGU3puBBXbnyPO5hSjoems8THhWkZlgOU4T4 MGPjpK0O3k1zaKax0UrUkT8CGAmIz2X2dojXDFTUPoAU5DQyx2TdDmspXFQfLmz2Hg43 jtvvad2YoVifvlBAE5u7327lxUKrfhmrBZu3jm66RwjjERaE6WWRXfA3iZD3OO8pTHdS DzdZjfixyZLmtvgc4AawZKqAfOzfF/5Dw2KPkPdgqjwztWhTPHMockzV+IN6wFCa86Q9 MjEqoZQilxLDkJ1+pOGoswUIanYuBMN/1CSDVk0n+yxlTpyCM7fcUfg5PXSUFkkBaT1K 8W8w== X-Gm-Message-State: AOJu0YyxTccrAvJXdCZzThqiHVQXFK81cDHzZYdc39jH9NCwiBcYQmoA 6fy2Pgu5Fuc/0F4PtoO6/C/qRb3gC/tj54IT8cE= X-Google-Smtp-Source: AGHT+IEVhZ3c0fRQafJ4wMYl4HHhOlCduqlnmi4uzO8Pu4kwPcRKFynV5tvpmKL+XXnnNprOppFbOF3q8HZ2eLIf0wQ= X-Received: by 2002:a7b:ca4a:0:b0:403:bb3:28c9 with SMTP id m10-20020a7bca4a000000b004030bb328c9mr1115796wml.38.1694579201813; Tue, 12 Sep 2023 21:26:41 -0700 (PDT) List-Id: Filesystems List-Archive: https://lists.freebsd.org/archives/freebsd-fs List-Help: List-Post: List-Subscribe: List-Unsubscribe: Sender: owner-freebsd-fs@freebsd.org MIME-Version: 1.0 References: In-Reply-To: From: Rich Date: Wed, 13 Sep 2023 00:26:29 -0400 Message-ID: Subject: Re: Could zfs use a gpu or x86 mmx instructions for speedup? To: Joe Mcguckin Cc: freebsd-fs@freebsd.org Content-Type: multipart/alternative; boundary="0000000000005e3a09060535f71c" X-Spamd-Bar: ---- X-Rspamd-Pre-Result: action=no action; module=replies; Message is reply to one we originated X-Spamd-Result: default: False [-4.00 / 15.00]; REPLY(-4.00)[]; ASN(0.00)[asn:15169, ipnet:2a00:1450::/32, country:US] X-Rspamd-Queue-Id: 4RlnTc45kCz3PlG --0000000000005e3a09060535f71c Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable It does use SSE/AVX for various things, which are MMX's bigger, badder siblings. Conceivably you could try using a GPU but you'd be a little hard-pressed to make it a win in some scenarios since you have a nontrivial overhead from round trips to the GPU and back, which you usually try to hide by doing lots of operations in one copy at once, but then latency might be exciting. You could also look at the discussion on OpenZFS around ZIA or the Intel QAT integration for gzip. - Rich On Wed, Sep 13, 2023 at 12:23=E2=80=AFAM Joe Mcguckin wrote: > Maybe for parity/checksum computation? > > --0000000000005e3a09060535f71c Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable
It does use SSE/AVX for various things, which are MMX'= s=C2=A0bigger, badder siblings.

Conceivably you could tr= y using a GPU but you'd be a little hard-pressed to make it a win in so= me scenarios since you have a nontrivial overhead from round trips to the G= PU and back, which you usually try to hide by doing lots of operations in o= ne copy at once, but then latency might be exciting.

You could also look at the discussion on OpenZFS around ZIA or the Intel= QAT integration for gzip.

- Rich

<= div class=3D"gmail_quote">
On Wed, Sep= 13, 2023 at 12:23=E2=80=AFAM Joe Mcguckin <joe@via.net> wrote:
Maybe for parity/checksum computation?

--0000000000005e3a09060535f71c--