PCI Speed and Overclocking: Test Results

 PCI Bus Overclocking Tests
 Motherboard
CPU
 Chipset  PCI Speed at 220 Base FSB Setting  Functional PCI Lock
Asus P4C800-E
Pentium 4 3.2C
Intel 875p 33.3 YES
Asus P4S800D-E
Pentium 4 3.2C
SiS 655TX 33.3 YES
Soltek SL-PT880PRO
Pentium 4 3.2C
VIA PT880 33.3 YES
AOpen AK86-L
Athlon 64 3200+
VIA K8T800 36.6 NO
ECS 755A2
Athlon 64 3200+
SiS 755 36.6 NO
Soyo CK8 Dragon Plus
Athlon 64 3200+
nForce3-150 36.6 NO

As you will see in results with the new VIA PT880 board for Intel, early testing raised questions about whether the PCI/AGP lock was working. However, tests with PC Geiger showed the new VIA PT880, as well as the Intel 875P and SiS 655TX, all have a working PCI lock. We have seen outstanding overclocking results with the Intel chipsets and SiS 655TX chipsets; the working PCI lock certainly contributes to that outstanding overclocking performance.

The big surprise here is that none of the current chipsets for Athlon 64 have a working PCI lock, whether they claim one or not. This was a complete surprise and our first inclination was to throw out the results, since speeds on the Athlon 64 are derived in a very different way with Hypertransport. However, oscilloscope results and tests elsewhere pretty much support our findings. There does not appear to be a working PCI lock on any current chipsets for Athlon 64. Certainly, this contributes to the poorer overclocking that we have seen on A64 inasmuch as on-board component frequencies and AGP bus are related to PCI bus.

Please keep in mind that we have only measured PCI bus speed with PC Geiger. While we expect on-board component speeds and AGP to be affected also, we cannot say that with certainty on Athlon 64 systems. Some reports on the web claim that AGP frequency is fixed on both VIA and nVidia chipsets on the Athlon 64. We can neither confirm nor deny these claims with the simple tests that we have performed here. Also keep in mind that results are for 6 individual motherboards. We can generally assume these results apply to the chipset used, but in some cases, this may be a stretch. The ECS 755A2 is designed to be a value board, for example, and 'PCI lock' may simply not have been implemented on this board. On VIA and nVidia chipsets for Athlon 64, our results do mirror those measured objectively at other sites using other VIA and nVidia A64 boards and we are more confident applying these results to chipset assumptions.

We will continue to look at PCI speeds in future tests of new boards and chipsets, particularly those designed for Athlon 64 processors. For now, it certainly appears that a floating PCI bus is one of the factors holding back overclocking on all current Chipsets for the Athlon 64.

PCI Speed and Overclocking: Test Configuration
POST A COMMENT

26 Comments

View All Comments

  • PrinceGaz - Monday, February 16, 2004 - link

    I'm still awaiting any sort of reponse to the first post in this thread about why no tests seem to have been done at around 233 or 234FSB to see if a 1/7 divider kicks in. Thats the first thing I'd have tried after establishing no PCI lock at say 210 or 220. And yes, you can lower the multiplier on A64 chips so they aren't an issue like I said. Reply
  • Pumpkinierre - Monday, February 16, 2004 - link

    Yes I second ionpro2's #13 suggestion. If you can get higher clockspeed by lowering the multiplier. try your geiger at these higher speeds and see what's going on. Reply
  • Icewind - Monday, February 16, 2004 - link

    Now im curious as well as how you got it that high Reply
  • ionpro2 - Monday, February 16, 2004 - link

    Wesley Fink: No way that the motherboard would've posted with no PCI lock. Think about it: the devices that are dangling off of the PCI bus on the board itself would be running *way* the heck out of specification at the overclocks achieved on the Aopen. Perhaps these have higher tolerences, but I seriously doubt the ALC655 would run at 43Mhz, or the Realtek Gigabit LAN chip, either. Most likely, there are merely windowed dividers for these chips, with a 1/6 and 1/7 ratio. It would make sense for this to kick in at 220Mhz -- otherwise, you are running the PCI out of spec almost 10% down; and I'd be willing to bet whatever optimization is done for aftermarket cards is done for higher PCI/AGP speeds, not lower. *Please* test the AOpen board at 9x240 and see what PCI speed is reported. Reply
  • Wesley Fink - Monday, February 16, 2004 - link

    Pumpkinierre -

    Your question is exactly the reason I stated that I don't know what is happening with AGP on A64 boards. With no PCI cards in my test setup, a working AGP lock could provide quite a bit of OC room. Some sites have reported AGP IS fixed on VIA, but I have no way of testing that. I can tell you for a fact that the PCI slots are not frequency locked on the A64 boards I tested, but the way frequencies are derived in HT it IS possible AGP is locked on some A64 boards.

    Reply
  • Pumpkinierre - Monday, February 16, 2004 - link

    I dont understand. You got 253MHzx8 and 240x9 on the Aopen A64 mobo when you tested it. That's a lot more than 220MHz. Even if the AGP was asynchronous, which I very much doubt, surely other components on the M'board would have seized?! Reply
  • impar - Monday, February 16, 2004 - link

    I would like to see the same test in nForce2 boards... Reply
  • Icewind - Monday, February 16, 2004 - link

    Sorry, guess I missed the defination of "go pound salt" so i'll take it with a grain of salt. Reply
  • idgaf13 - Monday, February 16, 2004 - link

    If all The motherboard chipset suppliers
    do the same thing ,such as the PCI lock,
    musi be a "deeper"reasoning.
    Certainly if possible ,contractually or technically, they would do it.
    They all need that special selling point ,A point of differintation.

    Basically "go pound salt" ,Icewind.
    Reply
  • Icewind - Monday, February 16, 2004 - link

    Cause VIA knows better and they still won't comply with the wishes of system builders. For that, I dispute them. Reply

Log in

Don't have an account? Sign up now