Received: from malur.postgresql.org ([217.196.149.56]) by arkaria.postgresql.org with esmtp (Exim 4.80) (envelope-from ) id 1ZE48V-0004NX-CO for pgsql-performance@arkaria.postgresql.org; Sat, 11 Jul 2015 23:24:47 +0000 Received: from localhost ([127.0.0.1] helo=postgresql.org) by malur.postgresql.org with smtp (Exim 4.84) (envelope-from ) id 1ZE48U-0000AF-JF for pgsql-performance@arkaria.postgresql.org; Sat, 11 Jul 2015 23:24:46 +0000 Received: from magus.postgresql.org ([2a02:c0:301:0:ffff::29]) by malur.postgresql.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_CBC_SHA384:256) (Exim 4.84) (envelope-from ) id 1ZE46v-0006n9-0K for pgsql-performance@postgresql.org; Sat, 11 Jul 2015 23:23:09 +0000 Received: from ore.jhcloos.com ([198.147.22.87]) by magus.postgresql.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_CBC_SHA384:256) (Exim 4.84) (envelope-from ) id 1ZE46r-0006Rq-4t for pgsql-performance@postgresql.org; Sat, 11 Jul 2015 23:23:07 +0000 Received: by ore.jhcloos.com (Postfix, from userid 10) id 3D4861E0C6; Sat, 11 Jul 2015 23:23:01 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=jhcloos.com; s=ore14; t=1436656981; bh=7qNpYQ+SbFimKjy5aOmSgT2Z/a/qP4NMOUqH/OdKvL4=; h=From:To:Cc:Subject:In-Reply-To:References:Date:From; b=DLxx5Dy3JH+h27NkCKsQG7OcNlfmwri2atTJYRDVnXzHnz4CFg1FB+wEH2GQC69Gp xWJ4UEOnJIMRHCe1554i8qxRbxs4SrBVq4Cm1kf+SLsBrvWo711sECvPsWmwq1MIw1 BPwHvfUmtkEvEYEPNF6ce4ZdcWEPkZPm62nqLwxI= Received: by carbon.jhcloos.org (Postfix, from userid 500) id 29C5C106FD888; Sat, 11 Jul 2015 23:18:21 +0000 (UTC) From: James Cloos To: "Graeme B. Bell" Cc: Andres Freund , Merlin Moncure , Craig James , "Joshua D. Drake" , postgres performance list Subject: Re: Hmmm... why does CPU-intensive pl/pgsql code parallelise so badly when queries parallelise fine? Anyone else seen this? In-Reply-To: (Graeme B. Bell's message of "Thu, 9 Jul 2015 08:59:26 +0000") References: <6E55113B-FA05-433C-9B21-623F798EE935@skogoglandskap.no> <559CB5BF.2080509@commandprompt.com> <20150708202733.GS10242@alap3.anarazel.de> User-Agent: Gnus/5.130014 (Ma Gnus v0.14) Emacs/25.0.50 (gnu/linux) Face: iVBORw0KGgoAAAANSUhEUgAAABAAAAAQAgMAAABinRfyAAAACVBMVEX///8ZGXBQKKnCrDQ3 AAAAJElEQVQImWNgQAAXzwQg4SKASgAlXIEEiwsSIYBEcLaAtMEAADJnB+kKcKioAAAAAElFTkSu QmCC Copyright: Copyright 2015 James Cloos OpenPGP: 0x997A9F17ED7DAEA6; url=https://jhcloos.com/public_key/0x997A9F17ED7DAEA6.asc OpenPGP-Fingerprint: E9E9 F828 61A4 6EA9 0F2B 63E7 997A 9F17 ED7D AEA6 Date: Sat, 11 Jul 2015 19:18:21 -0400 Message-ID: Lines: 16 MIME-Version: 1.0 Content-Type: text/plain X-Hashcash: 1:28:150711:graeme.bell@nibio.no::3O84QO9htwRZBMzo:0000000000000000000000000000000000000000AW3jl X-Hashcash: 1:28:150711:andres@anarazel.de::+AVuMy+or54b/9/0:000000000000000000000000000000000000000000fB9Hj X-Hashcash: 1:28:150711:mmoncure@gmail.com::/s6ba2ucdpx/LUVi:000000000000000000000000000000000000000000VrS74 X-Hashcash: 1:28:150711:cjames@emolecules.com::3szXDGb5g6CWUILq:0000000000000000000000000000000000000000aSdx X-Hashcash: 1:28:150711:jd@commandprompt.com::XehGSzWqKqUlaWLW:00000000000000000000000000000000000000005/S1G X-Hashcash: 1:28:150711:pgsql-performance@postgresql.org::4d7t5jmMp4kj8F5B:00000000000000000000000000005L7T/ X-Pg-Spam-Score: -2.6 (--) List-Archive: List-Help: List-ID: List-Owner: List-Post: List-Subscribe: List-Unsubscribe: X-Mailing-List: pgsql-performance Precedence: bulk Sender: pgsql-performance-owner@postgresql.org >>>>> "GBB" == Graeme B Bell writes: GBB> 1a. For example AMD CPUs list the number of integer cores (e.g. 16), GBB> but there is actually only half as many cores available for floating GBB> point work (8). So if your functions need to use floating point, your GBB> scaling will suffer badly on FP functions. That is half as many 256-bit float units; for scalar math and for 128-bit vector math each core gets a half of the float unit. Only for the 256-bit vector math do the schedulars have to compete for float unit access. -JimC -- James Cloos OpenPGP: 0x997A9F17ED7DAEA6 -- Sent via pgsql-performance mailing list (pgsql-performance@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-performance