Shuttle SB81P: Intel's 915G

by Evan Lieb on September 18, 2004 4:08 PM EST

Shuttle SB81P: Overclocking and Memory Testing


Overclocking Testbed
Processor: Pentium 4 3.4GHz LGA775
CPU Voltage: 1.3875V (default)
Cooling: ICE heatpipe cooling
Power Supply: Shuttle 350W PSU
Maximum OC:
(Default Ratio)
208FSB
3536MHz (+4.0%)
Maximum FSB 208FSB @ 14X multiplier

Unfortunately, Shuttle still has not found a remedy for Intel's overclock lock, something that has now been successfully breached by most retail S775 Prescott motherboards on the market. This is definitely something that needs to be resolved by Shuttle, as it would seem almost criminal for an SFF with such overclocking potential to not be able to reach near the overclocking boundaries set in BIOS. As you can see above, a 208MHz FSB overclock (4.0%) is underwhelming to say the least. Most retail S775 motherboards are getting over 260MHz FSB, with some reaching near 300MHz in special circumstances. Clearly, the SB81P isn't nearly in the same range.

Oddly enough, the Monarch Hornet Pro reached this exact same 208MHz FSB limit. We believe it's a coincidence, especially since the SB81P was able to run very smoothly at 214MHz FSB before rebooting later on after hours of stress testing. As a further reminder, we only shoot for the highest trouble-free overclocks possible in our reviews, which is why 214MHz FSB wasn't listed.

Memory Stress Test Results:

This memory stress test is very basic, as it simply tests the ability of the Shuttle SB81P to operate at its officially supported DDR memory frequency of 400MHz, at the best performing memory timings that our OCZ Platinum DDR400 modules will support. Memory stress testing was conducted by running RAM at 400MHz with 1 DIMM slot filled.

Stable DDR400 Timings
(2/2 DIMMs populated)
Clock Speed: 200MHz
Timing Mode: N/A
CAS Latency: 2.0
Bank Interleave: N/A
RAS to CAS Delay: 3T
RAS Precharge: 6T
Precharge Delay: 2T
Command Rate: N/A

While not the ideal 2-2-2-5 that we were looking for, 2-3-2-6 are plenty fast timings. In reality, memory timings or any other memory latency concerns really aren't terribly significant from a performance standpoint. They're nice, they're neat, and they do affect performance, but in the end they make very few visible differences. What you really should be concerned about is memory bandwidth and memory speed. That is if you want to overclock, as the SB81P doesn't "officially" support faster than 400MHz DDR speeds.

Shuttle SB81P: BIOS Shuttle SB81P: Performance Test Configuration
Comments Locked

6 Comments

View All Comments

  • aw - Wednesday, October 13, 2004 - link

    #5 - I too was wondering the same thing???
  • swturner - Sunday, October 3, 2004 - link

    Back on 9/18, the article ended with:

    "In fact, we're currently in the middle of testing an SFF just like that. We'll let you know how that SFF turns out later this week."

    So - how'd it turn out??
  • Ender17 - Sunday, September 19, 2004 - link

    after nForce4 is out, so several months
  • ViperV990 - Sunday, September 19, 2004 - link

    Any idea when an AMD-based P-chassis XPC will be out?
  • Evan Lieb - Sunday, September 19, 2004 - link

    Yes IceWindius, we'll have an SN95G5 review posted in a day or two.
  • IceWindius - Sunday, September 19, 2004 - link

    Nice system layout, but I really want it to be an Athlon 64 setup and not the Prescrap. You guys gonna review the new SN95G also?

Log in

Don't have an account? Sign up now