From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Keith Packard" Subject: Re: [PATCH] vulkan: Add VK_EXT_calibrated_timestamps extension (radv and anv) [v4] Date: Tue, 16 Oct 2018 14:51:41 -0700 Message-ID: <87woqh7e6q.fsf@keithp.com> References: <20181015230515.3695-1-keithp@keithp.com> <20181016053150.11453-1-keithp@keithp.com> <87bm7t8z3k.fsf@keithp.com> <8736t58usr.fsf@keithp.com> Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="===============0118021972==" Return-path: In-Reply-To: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" To: Jason Ekstrand Cc: ML mesa-dev , Maling list - DRI developers List-Id: dri-devel@lists.freedesktop.org --===============0118021972== Content-Type: multipart/signed; boundary="=-=-="; micalg=pgp-sha256; protocol="application/pgp-signature" --=-=-= Content-Type: text/plain Content-Transfer-Encoding: quoted-printable Jason Ekstrand writes: > You've got me almost convinced as well. Thanks for the diagrams! I think > instead of adding 1 perhaps what we want is > > max2(sample_interval_ns, gpu_tick_ns + monotonic_tick_ns) > > Where monotonic_tick_ns is maybe as low as 1. Am I following you correct= ly? Not quite; I was thinking that because the sample_interval_ns is measured by sampling the monotonic clock twice, the actual interval can be up to 1 tick longer, so max2(sample_interval_ns + monotonic_tick_ns, gpu_tick_ns) The gpu_tick_ns is computed, not measured, and so accurately reflects the maximum deviation in the measurements. =2D-=20 =2Dkeith --=-=-= Content-Type: application/pgp-signature; name="signature.asc" -----BEGIN PGP SIGNATURE----- iQIzBAEBCAAdFiEEw4O3eCVWE9/bQJ2R2yIaaQAAABEFAlvGXW0ACgkQ2yIaaQAA ABHIBw//cXKBxnyjIBy1/JZDi4U5DrRNEevWdsip3PW9aY3E3FFOj07oqR0QEK7L pD2i5ROGAG5A2264Pj3XXo4O6wWAJUFs+Jgyljn0HeYf2Egg1ftkMhairBnYaZqu rzMkpBXDO2MYpCSHVTAPCUa91ot3ugXwFWSYm+uGO3PUMbJtUQEHZOI2T0Q4IGYB J7gcMy0bYv1LuKmiCHWfCrnMkXv2M1swMh48/lCRrgQCKrsu/eFTcWi+ArgWLXO4 waGTePDuHmi18cF6+UdypPlxHucOLh7aLo6PsNQZmnVP0d1IRg/ykIfq0CXmniET uY4cGFD9FACrYK6Xl53IsW5OXqFyED9JKDh8qE5Bka1vZA/YdgLr7qsMMCj9Sfu/ Nyt5E5F0Eq5zqKHUqc59krH48cWSg9//39vxPTt11oS7mUxZhtvaBrsA0l+dUqUf tnymbnYEd9ou8xGKAtFe/evxCY46iGd33zBSNdxgPwx9asyBYo0oNwq2B3rcu4f2 8HFEltnTtCtw4GBWj/D3fXetdoh2rpRUKKSgkuhVXLrX82W0gaOsw9Or3WVaPF24 nMEa3QspOvKwDHvHQFTAB++lRTcjSsiOKGTaahMzVw29a3xRht+V4jcWIEWofdE5 Gao7ikkT+GLoxAy/aKHENj/pgCFGuCJxmPYtvEXVQMxF+AXt/Zo= =yglZ -----END PGP SIGNATURE----- --=-=-=-- --===============0118021972== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: base64 Content-Disposition: inline X19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX18KZHJpLWRldmVs IG1haWxpbmcgbGlzdApkcmktZGV2ZWxAbGlzdHMuZnJlZWRlc2t0b3Aub3JnCmh0dHBzOi8vbGlz dHMuZnJlZWRlc2t0b3Aub3JnL21haWxtYW4vbGlzdGluZm8vZHJpLWRldmVsCg== --===============0118021972==--