From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from anholt.net (anholt.net [50.246.234.109]) by gabe.freedesktop.org (Postfix) with ESMTP id 2E9E06E0FD for ; Tue, 27 Feb 2018 23:29:07 +0000 (UTC) From: Eric Anholt In-Reply-To: <20180227215040.11297-1-chris@chris-wilson.co.uk> References: <20180227215040.11297-1-chris@chris-wilson.co.uk> Date: Tue, 27 Feb 2018 15:29:04 -0800 Message-ID: <87muzut3jz.fsf@anholt.net> MIME-Version: 1.0 Subject: Re: [igt-dev] [PATCH igt] lib: Provide an accelerated routine for readback from WC List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: multipart/mixed; boundary="===============1807891086==" Errors-To: igt-dev-bounces@lists.freedesktop.org Sender: "igt-dev" To: Chris Wilson , igt-dev@lists.freedesktop.org List-ID: --===============1807891086== Content-Type: multipart/signed; boundary="=-=-="; micalg=pgp-sha512; protocol="application/pgp-signature" --=-=-= Content-Type: text/plain Chris Wilson writes: > Reading from WC is awfully slow as each access is uncached and so > performed synchronously, stalling for the memory load. x86 did introduce > some new instructions in SSE 4.1 to provide a small internal buffer to > accelerate reading back a cacheline at a time from uncached memory, for > this purpose. I think without a _mm_mfence() before the movntdqas, you can get stale results from movntdqa's little cache. --=-=-= Content-Type: application/pgp-signature; name="signature.asc" -----BEGIN PGP SIGNATURE----- iQIzBAEBCgAdFiEE/JuuFDWp9/ZkuCBXtdYpNtH8nugFAlqV6cAACgkQtdYpNtH8 nujMZQ//WoTmzbZZR1AqbyfJzYCFDODXT0Iz8GfNEZlzceoY7P1//7o6IN1IBmH0 7HtrWQ6NcOF+l6iINxUGd8jAppXAp3we8y5zTY27xhR4hJUfMLD3Sy/GsY9g95U9 bwGcvfSF9jCQGH8H3K0zkLAGdBErNbjEWJL/jSTD1IB9BL+KOq2fLQI/7rFNDgSG Hu5D5tuLD47ZPHIAsSVy/e9rRWBvwLSkKFZ/bO3u6D4wQ+Ws5rJSOJ34TWThEWeO xCLmt9TmlERuQzTW2dqh7IpTue/ZGcXfKvrswuqxpi/pTRb9bUSSjxNZCI+62NKl +iWmm7ApHfr93PNsOOZMM7huKX9zf6LriidJJLBKxOcoRC/j8Qu/Nsu25aGFADCR nohFL6l8jrf5kkcxRoLZWO9j9iBxlkx55PBibI8C060ajzlYYWu6RXSAR2St5ySd RmBZauAiV6FMFYx9ejTS343xyyd/RQEd+kyN3Vb5jGF++M9uT/QVXiLacVVMGPJ2 jDVQ/fasaN175kOoguncmTLS4sBx9x7BEE2+zdNBeOLaaDnMT/y+SpUVSlmFhNpk VBnfRb9OFukdwQLYPqGFJXhwvCqdrWDZ8PceGdkItQvdqXY5KtiX43g11mUprL2F YRdNZNits5qP+2eAFrGaEfTQB+2RVjh1kotbwQAjrR/XdUGh/TY= =rHtF -----END PGP SIGNATURE----- --=-=-=-- --===============1807891086== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: base64 Content-Disposition: inline X19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX18KaWd0LWRldiBt YWlsaW5nIGxpc3QKaWd0LWRldkBsaXN0cy5mcmVlZGVza3RvcC5vcmcKaHR0cHM6Ly9saXN0cy5m cmVlZGVza3RvcC5vcmcvbWFpbG1hbi9saXN0aW5mby9pZ3QtZGV2Cg== --===============1807891086==--