VOGONS

Search

Common searches

Asrock 4CoreDual SATA II / Windows 98 SE issues

Topic actions

First post, by paradigital

Posted on 
User metadata
Rank Oldbie
Rank
Oldbie

I've been building myself a benchmarking rig for older graphics cards, trying not to CPU constrain them so using LGA775 and a 3.2GHz Core 2 Duo (E4700 overclocked). Current spec is:

Dimastech Bench Table
ASRock 4CoreDual SATA II (Rev 1.x)
Intel Core 2 Duo E4700
4GB Crucial BallistiX Sport DDR2-800
Gainward BL:ISS 7800gs+ AGP
Integral V-Series 120GB SSD
Bitfenix Whisper M Series 750W 80-Plus PSU

It's a triple-boot system (so I can directly compare how OS choice affects performance) running on a 120GB SATA SSD (BIOS is in IDE mode for the SATA ports), Windows 98 SE on the primary partition, Windows 2000 Professional on the second, and Windows XP on the third.

In Windows 2000 and Windows XP, the system performs similarly well (within a few % of each other) in the only test I've run so far (3DMark2001SE) 35000~ marks.

Windows 98 SE on the other hand scores WELL UNDER that of 2k/XP (I don't believe 3DMark2001SE is multi-threaded?), coming in at around 18000~ marks (52% of the 2k and XP installs)

Now, I am using the re-released RLOWE memory patch so that Windows 98 plays nicely with the RAM (so I don't have to use different memory configurations potentially skewing the benchmark results), and Zak's modified 7-series driver for the GPU.

So first question: Does anyone know if 3DMark2001SE is multi-threaded? If it is then I guess a roughly 50% performance deficit makes sense.
Second question: For anyone who has previously used the RLOWE patch, has anyone seen a performance deficit vs simply removing physical RAM?
Third question: Has anyone used Zak's modified driver, and if so how did you find the performance?

Another issue that has presented itself is around the storage. If I try to bench using SiSoft Sandra 2001 then the storage benchmark never completes in 98 SE. I'm not sure if this is a quirk of using an SSD in 98, or something to do with drivers for the storage controller?

So fourth question: Does anyone have a good set of drivers for this motherboard on 98SE? Is there a particular version of the Via/Hyperion 4-in-1 I should use?
Fifth question: Does 98SE play nice with solid state disks or should I go back to using a mechanical drive?
Sixth question: Should I perhaps ditch the on-board SATA controller (even though it's in IDE mode)?

Finally, 98 SE seems unable to wake from sleep. If I leave the machine running until the display goes blank, there's no way I can wake it, keyboard and mouse input don't wake it, and neither does a brief press of the power button. The only course of action is a reset of a hard power-off.

Final question: Is this a driver problem with ACPI? Win2K and WinXP are happy to sleep and wake.

Reply 2 of 5, by VDNKh

Posted on 
User metadata
Rank Newbie
Rank
Newbie

I have the R2.0 version of your board and have the same issue. I've been searching Vogons, looking for others with the same issue, to try and piece together what's going on here. A comment on my build thread suggests it's an AGP issue.

First question: Not to my knowledge, a slower AMD CPU with the same GPU as me is scoring 5,000 points higher in 3DMark01.

Second question: There's no performance difference I saw with RLowe's memory patch

Third question: I haven't tried Zak's but I tried the Omega modded NVidia driver which actually made it worse.

Fourth question: I used 5.24A in 98SE and it's pretty stable. This thread tested a lot of chipset driver combos, all have the same problem.

Fifth and sixth question: I have an IDE SSD and it works fine, but SATA has no compatibility mode on R2.0 so it's apples to oranges for what your asking. 98 seems to like the RAID mode for SATA devices though, just make sure you don't run EMM386 while running a device in RAID mode.

Last question: I can't wake my machine from sleep either so I just don't use it.

Reply 3 of 5, by BitWrangler

Posted on 
User metadata
Rank l33t++
Rank
l33t++

Just a wildcard thought, wondering if the other cores aren't exactly twiddling their thumbs and are paging the AGP from time to time to see if they've got anything to do, thus interrupting the core that has got something to do and slowing things down.

Unicorn herding operations are proceeding, but all the totes of hens teeth and barrels of rocking horse poop give them plenty of hiding spots.

Reply 4 of 5, by VDNKh

Posted on 
User metadata
Rank Newbie
Rank
Newbie
BitWrangler wrote on 2021-12-06, 23:15:

Just a wildcard thought, wondering if the other cores aren't exactly twiddling their thumbs and are paging the AGP from time to time to see if they've got anything to do, thus interrupting the core that has got something to do and slowing things down.

Doesn't seem to be, adi88 was using a single core Athlon64 for his tests and is getting a similar slow down.

Reply 5 of 5, by TrashPanda

Posted on 
User metadata
Rank l33t
Rank
l33t

Could be something to do with these hybrid boards, the AGP on them isnt actually AGP but AGP emulated via the PCIe bus, WinXP doesnt have issues as it has native PCIe handling . .Win98 does not, so 98 is expecting the AGP bus to behave exactly like AGP should but the AGP via PCIe doesnt behave 100% like true AGP.

You are also using a 7600gs which is a native PCIe GPU using a bridge chip to provide AGP, try using a true AGP GPU and see if that helps .. the combination of emulated AGP and the bridge chip under 98 might not play nice.

I have a few of the Asrock 4core hybrid boards and they all have issues with 98 and the emulated AGP port on the boards, either it works perfectly or it works but I have weird issues with both drivers or programs not behaving correctly, different drivers also give different results and switching them around will generally cause further issues.

Go to top of page Go to top of page
Back to General Old Hardware
Close window
Close window