Date: Sat, 27 Nov 1999 14:27:55 -0700 From: Chris Wasser <cwasser@v-wave.com> To: Tom Embt <tom@embt.com> Cc: freebsd-questions@FreeBSD.ORG Subject: Re[2]: freebsd and ata-66 Message-ID: <4602.991127@v-wave.com> In-Reply-To: <3.0.3.32.19991127152202.0129edc0@mail.embt.com> References: <3.0.3.32.19991127152202.0129edc0@mail.embt.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Saturday, November 27, 1999, 1:22:02 PM, you wrote: > There should be a modest speed improvement with today's fastest drives > (like the IBM DPTA series, etc), but a drop in CPU overhead? I'd be > interested in seeing this measured (Just to be sure we're not comparing it > to wd driver PIO4 without all the flags turned on, or something). > I'm running a BP6 but haven't yet bought a drive for the ATA/66 controller. We did our own tests here on a BE-6, while they may not pass as "industry standard" tests, this is what we came up with: ABIT BE6 r1.00 / P2-400 / 128MB PC-100 Micron DFI P2XBL rD.1 / P2-400 / 128MB PC-100 Micron - Quantum Katana (KA) ATA-4 - Quantum Corona (CR) ATA-4 - Quantum ST UDMA33/ATA-2 - Quantum TM EIDE (PIO1-4) ----- ABIT board was using Highpoint-Tech HPT366 UDMA66 Controller (onboard) DFI was using Intel BX PCI-IDE controller (onboard) Quantum was chosen as they are the inventors of UDMA33/66 technology, thus making a good baseline choice. ----- First run through: Board Drive %CPU Sustained MB/s ========================================================= ABIT 18.2GB KA 7.4% 58.1 DFI 18.2GB KA 97.7% 7.1 ABIT 9.5GB CR 6.54% 58.4 DFI 9.5GB CR 96.4% 6.9 ABIT 6.4GB ST 6.89% 28.6 DFI 6.4GB ST 95.7% 5.8 ABIT 2.1GB EIDE 2.83% 13.9 DFI 2.1GB EIDE 90.5% 12.7 ========================================================= Each drive & board combination was tested 5 times, sequentially. The statistics for the HPT366 varied only on the CPU time, but never more then 0.2% upwards. The Intel-BX controller %CPU rose only 0.1% of first recorded value. As you can see even older drivers benefit from the ATA-66 interface, I think it's because on these ABIT boards the interface exists purely in PCI-space thus taking advantage of the PCI bus. The tests were conducted with a proprietary HDD benchmark from Ontrack (who's been around long enough that I'm satisfied they know what they're doing) Again, these are tests we conducted in-house in our labs but as I've seen the same results overall (give or take a percent or MB) with boards from Tyan, Soyo and ASUS .vs. the ABIT BE-6. The BP6 has the same controller so I think the tests should be valid in this context. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4602.991127>