VIA Apollo 133A AGP4x stuffs

[Oerg866]

Re: VIA Apollo 133A AGP4x stuffs

to unbury this thread


This issue was indeed a problem of chipset and mobo combo. My new board ASUS CUV4X-D has same problem wiht not automatically doing agp4x, but with rivatuner forcing agp4x mode, it works just fine!

[Shodan486]

Re: VIA Apollo 133A AGP4x stuffs

BTW guys, any updates on the progress in resolving the issue originating the thread? I'm merely interested in the Omega driver usage results, if it has helped so far...

[ratdude747]

Re: VIA Apollo 133A AGP4x stuffs

A Pentium M is basically just a P3 Tualatin in a package designed for laptops AND they go to higher clock speeds AND the mobos are USB2.0 and take DDR.
[It's a P3-Tualatin with all the P4 advantages.]
.

not to mention the Pentium m has SSE2, 1 or 2mb cache, and a nicer version of speedstep.

[PCBONEZ]

Re: VIA Apollo 133A AGP4x stuffs

There's an odd VIA mobo chipset [CLE266] that gives P3 USB2.0 and allows either/or DDR or PC133 RAM if the manufacturer wants to build the board that way.

Problem is VIA used it to push their EPIA line of CPUs so most boards that have that chipset have the sucky EPIA CPU soldered on. [No socket.]

I looked for one for a while but boards with that chipset and an actual socket 370 ZIF socket tend to be expensive for what they are.
It's more cost effective to find a MODT board that takes a Pentium M.
A Pentium M is basically just a P3 Tualatin in a package designed for laptops AND they go to higher clock speeds AND the mobos are USB2.0 and take DDR.
[It's a P3-Tualatin with all the P4 advantages.]
.

[PCBONEZ]

Re: VIA Apollo 133A AGP4x stuffs

I thought about that too - i have a PCI USB 2.0 card on mine, i bought it off my cousin for $10 about 3 years ago. And i also have a Silicon Image PCI SATA card + 320GB Western Digital HDD.

That's fine if you're happy with Windows.

I got sick of the add-in card method because hunting for Linux compatible USB 2.0 cards gets OLD real fast because most ads don't mention which USB chip they use.
That and when you get a good deal on a board but have to cough up an additional $10-$20 for an add-in card + shipping that pretty much kills the good deal aspect.

Never used Silicon Image cards for the same reason. [Linux]

.

[goodpsusearch]

Re: VIA Apollo 133A AGP4x stuffs

I see. Thanks for replying!

I have an Intel Pentium 3 Celeron Tualatin 1300MHZ @ 1456MHZ, 3xSDRAM 112MHZ 640MB, ATI All In Wonder Rage 128 Pro AGP, 2xHDD IDE 7200 20GB WD & 5400 40GB Samsung, BTC DVD ROM & Yamaha CD RW, lan PCI card and Soundblaster 128 PCI running windows XP service pack 3.

I use it mainly for TV and testing computer parts that I get for free.

It can be used for surfing the web too, only drawback is it can't play youtube videos without dropping frames..


I attach some photos I took of it 9 months ago.

[Th3_uN1Qu3]

Re: VIA Apollo 133A AGP4x stuffs

I only went past socket 370 because USB1 wasn't doing the job anymore when drives got big. [Too slow to transfer a whole drive.]

I thought about that too - i have a PCI USB 2.0 card on mine, i bought it off my cousin for $10 about 3 years ago. And i also have a Silicon Image PCI SATA card + 320GB Western Digital HDD.

@ goodpsusearch, it's mainly because they are extremely reliable. I've never had mine crash even once. And upgraded to the maximum (i only have ONE mod to go till i truly reach the limit of the system), you'd be surprised what a pair of PIII Tualatins can do. Just FIY, they keep up with low-end C2Ds at video encoding. It's not my main system, i have it at my grandparents' place.

I've grown accustomed to tinkering with that system every year, and it keeps on going so why get a new one? If i need an extra helping of modern tech i have two laptops bought last year. I bet that computer will outlive the two laptops though. I'm sure i'll be able to use it for at least 5 more years, and when it'll finally be truly obsolete i'll host a CS server on it or something.

[PCBONEZ]

Re: VIA Apollo 133A AGP4x stuffs

what do you find so fascinating with those Pentium 3 systems?

.
My can opener opens cans fast enough for my needs so should they come out with a faster one the upgrade would be pointless and a waste of money.
And, I just DON'T CARE how fast the neighbor's can opener is.

I only went past socket 370 because USB1 wasn't doing the job anymore when drives got big. [Too slow to transfer a whole drive.]
Even then I was still using one for office work up until last year.
It did IT'S job just fine.

[ratdude747]

Re: VIA Apollo 133A AGP4x stuffs

old, but has enough power to do heavier linux amd xp.

[goodpsusearch]

Re: VIA Apollo 133A AGP4x stuffs

what do you find so fascinating with those Pentium 3 systems?

[Th3_uN1Qu3]

Re: VIA Apollo 133A AGP4x stuffs

I have a dual PIII myself, but i'm running a BX chipset. Dual 1400/512k Tualatins on adapters, mobo needs modding (a new PLL IC) as it only does 100FSB thus the CPUs run at 1000. I said i'll mod it this year but i'm not too sure, patching in a new FSB clock is tricky business, and i sure as heck don't want to screw the board up.

Anyway, i run a Radeon 9800 Pro 256MB on that machine (last card compatible with AGP 2x). The card is overclocked and voltage modded, AND the mods resulted in gains. The 2x bus isn't limiting the card *that* much, so i'd leave it as it is.

Or, sell me the board. I'd be glad to have support for a HD4650 AGP, even if it would only run at 2x.

[ratdude747]

Re: VIA Apollo 133A AGP4x stuffs

no, come to think of it, my compaq presario 7000 does that too... same chipset, same problem with NVIDIA. or at least it did under linux with nvidia's drivers.

[Rulycat]

Re: VIA Apollo 133A AGP4x stuffs

Perhaps a silly question, but have you seen if there's a BIOS update for your board which fixes graphics card issues?

[RJARRRPCGP]

Re: VIA Apollo 133A AGP4x stuffs

I think the guy's not OCing anything, just getting it stable. Anyways it's true, if the firmware has a throttling option.

Just FYI, even though I saw that with a GeForce 4 Ti 4200. It would do that in 3D Mark 2001 SE and sometimes 3D Mark 2001 SE actually then gives an error, interestingly not from Windows, with a pop up from 3D Mark 2001 SE claiming it failed to init the graphics.

[Shodan486]

Re: VIA Apollo 133A AGP4x stuffs

That can also occur when a Nvidia GPU is overclocked too much.

I think the guy's not OCing anything, just getting it stable. Anyways it's true, if the firmware has a throttling option.

[RJARRRPCGP]

Re: VIA Apollo 133A AGP4x stuffs

I've tried RIVATuner to force it to AGP4x, which worked but anything 3D was extremely unstable (1 image.... wait 5 seconds... black screen..... new image) and lead to a locked up machine..

That can also occur when a Nvidia GPU is overclocked too much.

[Per Hansson]

Re: VIA Apollo 133A AGP4x stuffs

Some of the ATI 2xxx 3xxx 4xxx AGP cards support HDMI audio via a built in soundchip
This requires that the card gets two IRQ numbers, (one for the soundprocessor and one for the GPU) the BIOS must be coded for this...

I modded one such BIOS myself;
https://www.badcaps.net/forum/showth...highlight=hdmi
http://www.techpowerup.com/forums/sh...15#post1839815

[Shodan486]

Re: VIA Apollo 133A AGP4x stuffs

switch to ATI?

Aaaaah, my beloved Apollo133a - kay, let's get into this. First I gotta argue with ratdude - sorry pal, that ATI solution won't comply with this chipset - actually I experienced the same problems ONLY with ATI cards - the myth of Nvidia cards are not compatible with this particular chipset is crap - it does not depend on brands only - depends on the overall design of the board and its BIOS plus the VGA card itself.

During my tales with the Abit VP6, I tried almost every suitable (and powerful) VGAs - the most stable I had at avail (and the most overclocakble) was the 7600GS from Gainward - many caps, many MOSFETs used on it, not to mention I could have modded it, which I have seen, proved to be quite effective. Then I tried the 6800 - very same, the OC was about 50MHz lower (probably due older chip process). After that I received some ECC RAM and it proved to be even more OCable than non-ECC, so I pushed the AGP frequency a bit higher - exactly 86,6MHz. Even this level did not cause any trouble - just for fun tried something older, like MX400 and 440 - no difference, everything worked as supposed to. The only thing that bothered me was the inability of the NTuner to control the GPU fan speed properly...workaround was to set the fixed speed and the job's done.

Now I tried the ATI, coz they have that 4650 (i guess), which is the last of the AGP cards - but before I bought it I wanted to try any ati card, just for verification all's gonna be OK. And the torture started :

First - Not all VIA hyperion drivers were compatibile. But that was easily overcome - tried the drivers designed for KT400 - 600 & worked. And now, finally your problem - AGP speed...Not only we share the same issue, but additionally I had problems with the SideBand Adressing option and FastWrites/Reads:

I couldn't have both Fast Writes / Reads enabled, after the obligatory reset it always disabled the Fast Reads (the Catalyst), BUT (as I recall) it was capable running at 2x as well with SideBand enabled. This really surprised me, so many obstacles. So I'm trying, different drivers, different catalyst - nothings...then I reminded myself of Omega. This only fixed the sideband problem and the Fast R/W - the speed remained the same. End of story, really didn't spend much time testing it...just tried other cards, the only good one was the slow Radeon 9200, but that's from the Ice age.

I don't want to blame the VGA vendors, but try to install everything once again, try the newest Hyperion drivers and XP SP2, then try some Nvidia first if it is possible - just give it a try...if not, try the ATI and use Omega drivers, maybe you'll have some luck.

Just to summarize the whole problem, any ideas are welcome: Why the Abit VP6 has no problems with Nvidias and many problems with ATI. And why does it not have problems with R9200? And and and...

I think ALL factors, from proper power management to the AGP slot / VGA card itself through ideal BIOS settings to the end of compatible OS + drivers installation - should be considered crucial towards solving such issues.

Last thing that comes to my mind is the AGP driving value - 7600GS was happy with the CD value, on the other hand 7950 required EA to run without any problems/glitches/falls. This setting increases the signal potential of the graphic subsystem, simply put it tries to ensure stability. Don't worry, you won't fry anything, had the FF value many times set and all is cool.

[Per Hansson]

Re: VIA Apollo 133A AGP4x stuffs

Try enabling AGP 4x but disabling AGP fastwrites, might make it more stable

[ratdude747]

Re: VIA Apollo 133A AGP4x stuffs

switch to ATI?