Ok fellas, those of you with the 8KHA+ may want to have a look:
First up, system specs:
InWin Q500 24" full tower case w/2 Sunon 80mm case fans
EPoX 8KHA+ (BIOS date 3/14/02)
AMD Athlon XP 1600+ (10.5 x 133 default, voltage bumped up to 1.80 from 1.75)
2x256 Generic Samsung RAM (RAM setting in BIOS "Normal", voltage bumped up to 2.70 from 2.50)
ATi Radeon 8500 AGP retail (275/275 stock)
Enermax 431W Whisper PSU
SB Live! Value (PCI Slot 2)
D-Link 528 & 538 10/100 NIC's (PCI Slots 3&4 respectively)
Maxtor 40Gb 7200RPM ATA100 HDD (Single NTFS partition, alone on primary channel)
Toshiba 16x DVD (Secondary master)
LG 8120B CD-RW (Secondary slave)
Logitech USB iFeel optical mouse
Logitech USB Rumblepad
Microsoft USB GameVoice
Second, the story:
Okay, so we've all had some level difficulty getting these boards stable with newer video cards.
People have made mention of everything from bad 5V rails to poor 4x implementation to Southbridge
incompatibility with the SB Live! series. Suffice to say, after reading post after post, and getting loop
after loop, I was extremely frustrated. I decided on a whim to call ATi and find out what the deal was,
and after speaking with one of the driver developers, he mentioned to me there were a few important
things to make sure of when installing the new Radeon cards.
Specifically,
1. The AGP aperture MUST be set to 64MB or the drivers would not install properly.
2. The VIA 4-in-1 drivers are better installed in safe mode to avoid other device conflicts.
Now after doing so, I saw a marginal improvement in terms of stability, but nothing to really write home about.
I still looped, however I usually got a little further into 3dMark before the loop would occur.
I monitored and logged the power usage on all lines at a second-by-second basis to ensure it
was not a issue with the PSU. All lines reported excellent power output, including
+12V - 12.6 avg, lowest 12.48.
+5V - 4.99 avg, lowest 4.92.
+3.3V - 3.36 avg, lowest 3.31
VCore - 1.75 no variation
VIO - 1.54 avg, lowest 1.52
VDIMM - 2.50 no variation
So why the hell is this thing still looping? Time for some hard registry searching :-(
After about 40 minutes of trying to decipher Microsoft's intelligently laid-out registry,
I came across something very interesting.
There was a key value which read "EnableVia4x" with a value of "1".
Obviously, this showed that 4x was in fact enabled in the drivers.
Many people reported excellent stability after backing the AGP clock to 2x.
So logically, setting the above value to zero should disable the 4x instability, correct?
Reboot, re-run 3DMark....LOOP! ARGH! Unbelievable...at this point I'm ready to take the 9mm to
the whole shebang..but wait a second...DUH! I forgot to turn AGP down to 2x in the BIOS!!
Reboot, re-run...AND it finishes the TEST!! With an excellent score of 8138 3DMarks!!
Well once through proves nothing right? Ok, I run it in a repetitious loop 15 times...
Surprise! Infinite loops!!!
After all that effort, there's still something not quite right here....
Third, The Last Ditch Effort:
Well since there's been so much talk about voltages relating to stability recently, on a wing
and a prayer, I decide to try one last idea. Bump the CPU Core voltage to 1.80 from 1.75,
and bump the VDIMM voltage to 2.7 from 2.5. Boot up XP and re-run 3DMark......
BAM!! 8204 3DMarks...Once again, one test through proves nothing to me, so let's run it
in a loop 15 times and walk away from the box till it's done.
(Maybe "A watched pot never boils" is the status quo here...)
Imagine my amazement to come back an hour later and see a score of 8212 on the screen.
I must be on glue, did that really finish without dying? It did, and admirably at that.
Ok, so DirectX seems stable, what about OpenGL?
Let's Install Q3 and Q3Bench and test.
I set Q3Bench to run all tests repeatedly, at every setting I could think of, and low and behold,
the system stood up to Q3 with grace.
Finally, the tweaks:
Well, I'm never one to leave well enough alone, so now's the best time to try and break something,
while I'm still feeling good about myself.
Back to the BIOS, set the RAM timings to "turbo", boost the bus right up to 160 MHz off the bat.
(No sense in small increments right, I need more POWER!)
XP boots up, I run 3DMark another 15 times (just short of 9000 3DMarks), and I'll be damned
if the system didn't stand up like a champion. Q3? GLExcess? Solid as a stone wall. I'm not big
into overclocking so I knocked the bus back down to 133 and left the memory timings at normal,
just to ensure things went as smooth as possible.
If anybody else happens to have the same/similar setup and would like assistance, please let me know
via email or ICQ (check my profile) and I'll be happy to help you out!
Regards,
SmoothVision
First up, system specs:
InWin Q500 24" full tower case w/2 Sunon 80mm case fans
EPoX 8KHA+ (BIOS date 3/14/02)
AMD Athlon XP 1600+ (10.5 x 133 default, voltage bumped up to 1.80 from 1.75)
2x256 Generic Samsung RAM (RAM setting in BIOS "Normal", voltage bumped up to 2.70 from 2.50)
ATi Radeon 8500 AGP retail (275/275 stock)
Enermax 431W Whisper PSU
SB Live! Value (PCI Slot 2)
D-Link 528 & 538 10/100 NIC's (PCI Slots 3&4 respectively)
Maxtor 40Gb 7200RPM ATA100 HDD (Single NTFS partition, alone on primary channel)
Toshiba 16x DVD (Secondary master)
LG 8120B CD-RW (Secondary slave)
Logitech USB iFeel optical mouse
Logitech USB Rumblepad
Microsoft USB GameVoice
Second, the story:
Okay, so we've all had some level difficulty getting these boards stable with newer video cards.
People have made mention of everything from bad 5V rails to poor 4x implementation to Southbridge
incompatibility with the SB Live! series. Suffice to say, after reading post after post, and getting loop
after loop, I was extremely frustrated. I decided on a whim to call ATi and find out what the deal was,
and after speaking with one of the driver developers, he mentioned to me there were a few important
things to make sure of when installing the new Radeon cards.
Specifically,
1. The AGP aperture MUST be set to 64MB or the drivers would not install properly.
2. The VIA 4-in-1 drivers are better installed in safe mode to avoid other device conflicts.
Now after doing so, I saw a marginal improvement in terms of stability, but nothing to really write home about.
I still looped, however I usually got a little further into 3dMark before the loop would occur.
I monitored and logged the power usage on all lines at a second-by-second basis to ensure it
was not a issue with the PSU. All lines reported excellent power output, including
+12V - 12.6 avg, lowest 12.48.
+5V - 4.99 avg, lowest 4.92.
+3.3V - 3.36 avg, lowest 3.31
VCore - 1.75 no variation
VIO - 1.54 avg, lowest 1.52
VDIMM - 2.50 no variation
So why the hell is this thing still looping? Time for some hard registry searching :-(
After about 40 minutes of trying to decipher Microsoft's intelligently laid-out registry,
I came across something very interesting.
There was a key value which read "EnableVia4x" with a value of "1".
Obviously, this showed that 4x was in fact enabled in the drivers.
Many people reported excellent stability after backing the AGP clock to 2x.
So logically, setting the above value to zero should disable the 4x instability, correct?
Reboot, re-run 3DMark....LOOP! ARGH! Unbelievable...at this point I'm ready to take the 9mm to
the whole shebang..but wait a second...DUH! I forgot to turn AGP down to 2x in the BIOS!!
Reboot, re-run...AND it finishes the TEST!! With an excellent score of 8138 3DMarks!!
Well once through proves nothing right? Ok, I run it in a repetitious loop 15 times...
Surprise! Infinite loops!!!
After all that effort, there's still something not quite right here....
Third, The Last Ditch Effort:
Well since there's been so much talk about voltages relating to stability recently, on a wing
and a prayer, I decide to try one last idea. Bump the CPU Core voltage to 1.80 from 1.75,
and bump the VDIMM voltage to 2.7 from 2.5. Boot up XP and re-run 3DMark......
BAM!! 8204 3DMarks...Once again, one test through proves nothing to me, so let's run it
in a loop 15 times and walk away from the box till it's done.
(Maybe "A watched pot never boils" is the status quo here...)
Imagine my amazement to come back an hour later and see a score of 8212 on the screen.
I must be on glue, did that really finish without dying? It did, and admirably at that.
Ok, so DirectX seems stable, what about OpenGL?
Let's Install Q3 and Q3Bench and test.
I set Q3Bench to run all tests repeatedly, at every setting I could think of, and low and behold,
the system stood up to Q3 with grace.
Finally, the tweaks:
Well, I'm never one to leave well enough alone, so now's the best time to try and break something,
while I'm still feeling good about myself.
Back to the BIOS, set the RAM timings to "turbo", boost the bus right up to 160 MHz off the bat.
(No sense in small increments right, I need more POWER!)
XP boots up, I run 3DMark another 15 times (just short of 9000 3DMarks), and I'll be damned
if the system didn't stand up like a champion. Q3? GLExcess? Solid as a stone wall. I'm not big
into overclocking so I knocked the bus back down to 133 and left the memory timings at normal,
just to ensure things went as smooth as possible.
If anybody else happens to have the same/similar setup and would like assistance, please let me know
via email or ICQ (check my profile) and I'll be happy to help you out!
Regards,
SmoothVision
Facebook
Twitter