Date: Fri, 1 Feb 2002 22:50:06 +0100 From: Erik Trulsson <ertr1013@student.uu.se> To: =?iso-8859-1?Q?G=E9rard?= Roudier <groudier@free.fr> Cc: Jason Evans <jasone@canonware.com>, Kenneth Culver <culverk@yumyumyum.org>, "Cameron, Frank" <Cameron@ctc.com>, Terry Lambert <tlambert2@mindspring.com>, David Malone <dwmalone@maths.tcd.ie>, "'freebsd-current@freebsd.org'" <freebsd-current@FreeBSD.ORG> Subject: Re: AMD AGP Bug Message-ID: <20020201215006.GA50090@student.uu.se> In-Reply-To: <20020131211810.B1769-100000@gerard> References: <20020131173408.B63502@canonware.com> <20020131211810.B1769-100000@gerard>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, Jan 31, 2002 at 09:32:48PM +0100, G=E9rard Roudier wrote: >=20 >=20 > On Thu, 31 Jan 2002, Jason Evans wrote: >=20 > > On Wed, Jan 30, 2002 at 11:14:48PM +0100, G=E9rard Roudier wrote: > > > > > > Linux can be fixed, but the useless writes of the existing Athlons fr= om > > > the very fast cache to the relatively very slow memory cannot. And all > > > Athlon users may well pay this penalty under any OS... unless we wan= t to > > > disable caching. :) > > > > Have you done benchmarks to show that the speculative writes are useless > > often enough to cause enough memory bus contention that overall perform= ance > > is degraded, despite the speedups when the speculative writes are valid? >=20 > I haven't done any benchmark of this sort, neither intend to do any since > I haven't time for that. But I wrote in my email that my 2 Athlon systems > worked fine and fast, just to indicate that for normal use I didn't see > any performance problem at all. >=20 > > I > > suspect that AMD in fact performed such tests; otherwise they wouldn't = have > > gone to the trouble of implementing speculative writes. >=20 > The Athlon rewriting same value to cacheable memory under the knees of > programmers looks a severe issue to me if it is true. Not only AGP memory > can be affected. What about SMP, MMIO (if some cacheable mapping exists), > etc...? I am not familiar with the acronym MMIO is so I can't comment on that.=20 In general though, having memoryspace used for memory-mapped I/O devices (including AGP) marked as cacheable is a bad idea unless you are very careful and know exactly what you are doing. For SMP it shouldn't be any problem. Multi-CPU systems normally run some cache-coherence protocol between themselves to make sure that things like this is not a problem. >=20 > In my opinion, OSes having some cacheable mapping to AGP memory is not the > real problem. Just it has revealed the AMD issue. It might be argued that there should be some cache-coherence protocol between the CPU and the AGP device. Not knowing how AGP is specified I don't know if this interaction between the CPU and AGP is a bug or just working as specified. I suspect it is the latter though. --=20 <Insert your favourite quote here.> Erik Trulsson ertr1013@student.uu.se To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20020201215006.GA50090>