Anyone with Asus P3V4X experience

[dwright]

P3V4X will be available here in the Toronto area in a few days. Has anyone had experience with this board? I can't find any reviews of it; it might be too new. I downloaded the manual from Asus and the pcb layout and features looks similar to the P3B-F but using the latest Via chip-set Apollo Pro 133A instead of Intel BX. I have a P500e FC that clocks at 667 at the standard 1.6v (I haven't tried for anything higher). The P3V4X would give me the correct agp setting for the G400max.
What would I need to do above the normal install of win98se and win2000 retail (like install VIA drivers)?
Thanks

[RoGuE]

I've got an Asus P3V4X at the moment (evaluating it). Generally I'm impressed with it. As you say the layout of the board is almost identical to the P3B-F. The board has worked faultlessly with both a Celeron 366@550, and an FC-PGA PIII 500E @ 667MHz. It also has AGP 4X, but don't expect any kind of performance gain over 2X.

There are drivers on the supplied CD for Win98SE, not sure about Win2K.

Chris.

[dwright]

The P3V4X arrived late last week. I swapped out the P3B-f. The VIA Apollo 133a chipset seems be ok and the g400max operates fine with it and at agp 2x. The bios is a beta so some hardware functions like ultra 66 have not been enabled yet.
Win2000 works perfectly on this new chipset, I didn't even have to load any VIA drivers. Win98se was another story, only one older version of via drivers (on the cd) gives good performance (according to the reviews). I hope a bios update fixes the problem. I can run my PIII500e at 750 (150 fsb with the default voltage). I normally run it at 667.
Video problems occur in win98se with the latest tgl130 driver and QIII. There is nothing but a white flashing screen. I get out of it by hitting the web home key that gets me back to a the desktop where I restore the 32 bit colour and 1280*1024 desktop. The full icd is ok once I recovered from opengl lockups by setting agp memory to 256mb.

[MaXi]

Hi, I've also just bought me a P3V4X, and jumped it up to 133MHz at the first run with a PIII650E (100). The AGP was set to 256MB and 2xAGP... In 3DMark2000 is froze...

I went to opgrade the bios to 102 but I ****ed it op and had to get it flashed all over again...

What are you're AGP settings and other settings... What about the software, there's some AGP software also, what have u done there?

Fast specs are:
ASUS P3V4X
Pentium 650E Box
256MB 133MHz Ram
SBLive!
and so on...

[paulcs]

I bought one as well for a review I'm writing. I've done a fair bit of research on the board, both on USENET and the web. I have a couple of immediate concerns.

The VIA drivers that you have to use to avoid the huge 3D performance hit, 4.17, appear to have never have made it to VIA's site. (I've seen reports claiming they're very similar to 4.16.) Users of other Apollo Pro 133A boards have claimed they can't install the 4.17 drivers.

This *suggests* that they are proprietary drivers, written by VIA or modified by Asus for the P3V4X. That could mean long waits inbetween driver releases or worse. VIA has released three sets of drivers since 4.17, and none of them function properly with the board. Asus has posted v. 4.20 despite the fact that they have been the cause of the 40-50 fps loss in Quake3 reported on Tom Pabst's site and confirmed by Anand.

Asus has been approached about this, and they have yet to respond as far as I know. Hopefully, the popularity of the board will encourage Asus and VIA to provide regular updates.

Paul
paulcs@flashcom.net

[MaXi]

Isn't more the BIOS than the drivers?? The first benchmarks with the ASUS board was with the first BIOS, the 101 revision. Then a beta BIOS of 102 came out, and wooops, the frame/sec went up... Now there's a final 102. Well, I haven't tried to play with drivers yet, so...

[paulcs]

There is an issue with the drivers. Both Tom Pabst and Anand reviewed the board. In both cases, they downloaded and installed the latest VIA drivers, which is natural. I would have done the same thing. Both noted that the PV34X was significantly slower in certain areas than other Apollo Pro 133A boards.

Anand had seen similar behavior in, if memory serves me correctly, a board based on the older Apollo Pro 133 (no "A") chipset. This was corrected with a BIOS upgrade, which changed a parameter that increased stability but slowed the board down. Anand assumed the problem was the same with the P3V4X, and, I think, that's where the BIOS theory originated.

After trying the new VIA 4.20 drivers, Tom Pabst downgraded to the drivers that shipped with the board, 4.17. He reported that the speed problems, and they are really bad, were corrected by using the shipping drivers.

Anand heard about this, reran his tests, and confirmed Tom's findings. Since then, others have tried the VIA 4.18 and 4.19 drivers and reported similar slowdowns. I appears the 4.17 drivers are the only fix to date.

Paul
paulcs@flashcom.net

[dwright]

Hi Paulcs,
Please let me know if ACPI is enabled in win98se. I reformatted and installed win98se and ACPI is enabled by default. Do you think it results in lower performance? With my P3B-F ACPI is not enabled its using APM.
Win2000 has better performance with the P3V4X than win98se on this system. With the P3B-F system win98se is best.
Thanks

[paulcs]

I haven't installed the board yet. I am in the process of finishing up tests on the BX board I'm comparing it with for an article I'm going to write.

The first think I would do is check to see if there is an option to disable APCI in the BIOS. (I bet this was the first thing you did.) If not, we might be waiting for a BIOS upgrade. Check the Power Management section of the BIOS and see if there's a way to disable it.

I've heard the Windows 2000 drivers are pretty good and there is no need to upgrade them. Are you using the 4.17 drivers? Have you upgraded to BIOS ver. 1002? It recently had the beta designation removed.

Paul
paulcs@flashcom.net

[paulcs]

The funny thing is, while some P3V4X users are annoyed about being forced to use old drivers, some people who own other Apollo Pro 133A boards have been going crazy looking for them. I guess some people didn't read Tom's and Anand's reviews carefully, and now they think these drivers, which apparently are similar to the 4.16 4-in-1's, are some sort of magical cure for issues with their non-Asus boards.

Until today, when they started popping up all over the place, nobody could find them because they were only to be found on our installation CD's. This, I believe, made them even more attractive.

Paul
paulcs@flashcom.net

[dwright]

HI Paulcs,
I am at bios 1003.001 beta released today. 1002 final was installed for 24 hours. They do not have an option to disable ACPI (that I can find). On bx boards it takes extra effort to enable ACPI! I did install 4.17 AGP and verified the install by checking the registry and the version of the drivers in windows system. I don't think the ide drivers are install though even though I tried.
I ran into a blue screen when running hdtach in win98se. It recovers but the ms explorer usb mouse slows down to a near stop (I have usb ms-kb, ms-js, printer, monitor and mouse). This does not happen in win2000 where everything is working. It's like this board and g400max are made for win2000 (no via drivers required, they are built-in).
In win2000 I had to run 3dmark2000 several times to get a result (beats win98se by 200+ marks). 3DMark2000 returns to desktop much of the time, with no results. I think this is a 3dmark2000 problem because the gefore on the P3B-F does the same thing.

[paulcs]

I installed the P3V4X yesterday, and you're right. I couldn't find any obvious way to disable ACPI.

It was kind of an ugly install too. The machine wouldn't post in jumperless mode. I then switched to "jumper" mode and *carefully* set the jumpers for a PIII 600E. It wouldn't boot. I got a continous series of long beeps instead. I then turn all the dip switches to off, which is where they were in the first place, and it booted at 800 MHz (6x133). It's a good thing my CPU can boot into DOS at this speed, because I couldn't get it to post at any other.

I couldn't get it to run in at any speed but 800 MHz until I upgraded the BIOS to v. 1002.

I run SETI clients in a DOS window. I can run two of them simultaneously at 800 MHz without a hitch, but the machine locks up hard when I close them in the conventional manner. I can avoid this by shutting them down with ctrl-alt-delete. I have shut down errors galore as well.

I haven't tested USB yet. I'm going to reinstall Windows 98 tonight and see what I can do with this thing.

By the way, it's memory benchmarks in Sandra 2000 were low compared to marks I got with a BX board and the same configuration.

Paul
paulcs@flashcom.net

[This message has been edited by paulcs (edited 10 March 2000).]

[dwright]

Reloading windows is the way to go. When I first installed the board it tried it without reloading and received many blue screens and hangs.
The first time I reloaded I botched up trying to get rid of ACPI so yesterday I formatted and installed win98se.
Once you load the matrox drivers and setup the desktop it hangs on simble stuff and won't shutdown. To fix it go to default vga, and install the drivers from the ASUS cd (I checked off only the first two ide and agp). Reboot and check the install. On my system the disk dma became enabled but I don't think it installed via ide drivers. I then reloaded the matrox 5.52 drivers. I also loaded the update from MS to fix shutdown problems (I don't think this was the cause though).
One remaining problem is that hd-tach gives a blue screen (___.vxd error when I run it). It is recoverable though and migh be related to the ide driver install. This does not happen in win2000.
I tested several versions of Matrox drivers with 3dmark2000 (scores 5.30 = 2731 but gives weird colours, 5.41 = 2597, 5.50 = 2779, 5.52 = 2790). I did not install turbogl it gives a flashing white screen in previous tests.
I have a feeling that ACPI is causing some loss in the performance, but I can't prove it because I cant get rid ACPI. Win2000 gives better test scores.
My processor P500E @667 (ASUS socket 370-133), sblive, nic, tv-wonder, PT795,and all i/o is USB and all devices seem to be ok so far).

[dwright]

Sandra memory bench mark results:
cpu/memory 306 mb/s
fpu/memory 310 mb/s
Does this compare with yours?

[paulcs]

I'm doing worse. I found this benchmark to be all over the place, and I get fairly different results every time I run it. Here are three runs:

cpu/memory 271 mb/s
fpu/memory 309 mb/s

cpu/memory 276 mb/s
fpu/memory 322 mb/s

cpu/memory 292 mb/s
fpu/memory 327 mb/s

With the exact same components and a BX board, I got much higher marks:

cpu/memory 344 mb/s
fpu/memory 380 mb/s

cpu/memory 356 mb/s
fpu/memory 403 mb/s

cpu/memory 340 mb/s
fpu/memory 384 mb/s

Paul
paulcs@flashcom.net