eva2000
07-30-2003, 03:49 AM
4x 256MB ADATA PC4000 single sided memory rated for 3-4-4-8
tested on
Albatron PX865PE PRO2 @ 2.85v vdimm
Abit IC7 @ 2.8v vdimm (2.76v real idle vdimm)
==============================================
2.4C
MCX4000 rev1 + 120x35mm Sunon 120cfm fan + TT fan speed controller
128MB Gigacube 9600PRO at default
120GB WD 1200JB
Abit IC7 15 bios
vdimm = 2.8v bios (real 2.76v idle voltage)
vAGP = 1.6v
AGP/PCI: Fixed 66/33
NB Strap: 800FSB
HT enabled
Game Accelerator: Auto
Refresh Cycle Time: Auto
Read Delay Adjust: Auto
Read Delay: Auto
Command Per Clock: Auto
PCI Latency: 64
SATA: disabled
Firewire: disabled
==============================================
2.4C
MCX4000 rev2 + 92x38mm Sunon 120cfm fan + TT fan speed controller
128MB Gigacube 9600PRO VIVO at default
120GB WD 1200JB
Albatron PX865PE PRO2 1.02 beta bios
vdimm = 2.85v bios
vAGP = 1.6v
AGP/PCI: Fixed 66/33
Enhanced Performance disabled
Extreme timings disabled
OC Stability DDRAM disabled/blanked out
HT enabled
Spread spectrum disabled
Promise IDE Raid disabled
SATA controller disabled
Voice Genie enabled
==============================================
notes:
1. Single and 2 stick max oc's have been so far around 274-277mhz at 2.85v on PX865PE PRO2 for memtest86 upto 75mins looping test #5 and #6. I'll be revisiting single and 2 stick ocs and memtest86 stability tests in detail later on.
2. Introducing 4 sticks sees max oc's drop considerably on both PX865PE PRO2 and IC7. I've memtest86 tested 4 sticks looping test #5 and #6 for upto 24+ hrs still haven't found the absolute fault free speed for 4 sticks as yet.
==============================================
Memtesting results (4x 256MB ADATA PC4000 @ 3-4-4-8):
So far I've tested memtest86 looping test #5 and #6 for upto 24+hrs and for 4x 256MB ADATA PC4000, oc'ing headroom is way down for 2.76-2.85v vdimm.
---------------------------------
Abit IC7:
Come to conclusion 4 sticks on IC7 won't work due to lack of vcore. Under load in windows vdimm at 2.8v in bios drops to as low as 2.75v. Managed to do some benchies at 233, 250 and 255mhz even though memtest86 errored out.
Memtesting 233mhz 3-4-4-8 at 2.8v (2.76v real idle voltage) failed, after 237mhz 3-4-4-8 failed memtest86 with 18 errors after the 191st pass of test #5 loop, approximately 3hrs 11+ mins into test.
IC7's bios limit doesn't allow for >255mhz 1:1 as the pc boot screen freezes at memory display part, or memtest86 disk doesn't boot and stays at boot screen.
255mhz 3-4-4-8 AUTO memtest86 bandwidth = 2920MB/s failed 1st pass with 4 errors
250mhz 3-4-4-8 AUTO memtest86 bandwidth = 2866MB/s failed 5th pass with 12 errors
248mhz 3-4-4-8 AUTO memtest86 bandwidth = 2838MB/s failed 11th pass with 4 errors
246mhz 3-4-4-8 AUTO memtest86 bandwidth = 2818MB/s failed 32nd pass with 16 errors
245mhz 3-4-4-8 AUTO memtest86 bandwidth = 2804MB/s failed 13th pass with 12 errors
243mhz 3-4-4-8 AUTO memtest86 bandwidth = 2784MB/s failed 10th pass with 12 errors
240mhz 3-4-4-8 AUTO memtest86 bandwidth = 2750MB/s failed 214th pass with 12 errors
237mhz 3-4-4-8 AUTO memtest86 bandwidth = 2715MB/s failed 191st pass with 18 errors
233mhz 3-4-4-8 AUTO memtest86 bandwidth = 2668MB/s failed 31st pass with 30 errors [weird more errors sooner]
200mhz 2.5-3-3-5 Turbo memtest86 bandwidth = 2539MB/s failed 1st pass with 1920 errors
---------------------------------
Albatron PX865PE PRO2:
Currently memtesting at 245mhz 3-4-4-8 at 2.85v, seems I'm getting strange memtest results with 4 sticks under 255mhz, errors are showing up quicker than when > 255mhz :confused: Managed to do some benchies at 255, 260mhz and 270mhz even though memtest86 errored out.
Memtesting 255mhz 3-4-4-8 at 2.85v, after 260mhz 3-4-4-8 frooze in 3dmark2003 and failed in SuperPi 16m test. Only memtest looped test #5 for 228 passes error free before stopping it myself. Probably, if i fan it for 24hrs it would of shown errors at 260mhz 3-4-4-8.
PX865PE PRO2's 2.85v vdimm helped alot more in holding some decent speed for 4x 256MB ADATA PC4000 sticks of memory compared to IC7. Unfortunately, PX865PE PRO2 has a max vcore of 1.6v (undervolts to 1.52v), but for 4 sticks the oc'ing headroom is limited to well under what 2.4C's max oc is for 1.52-1.6v vcore.
275mhz 3-4-4-8 memtest86 bandwidth = 3145MB/s failed memtest86 immediately
270mhz 3-4-4-8 memtest86 bandwidth = 3088MB/s failed memtest86 test#5, 1st pass with 10 errors
268mhz 3-4-4-8 memtest86 bandwidth = 3065MB/s failed test#5, 1st pass with 2 errors
267mhz 3-4-4-8 memtest86 bandwidth = 3054MB/s failed test#5, 1st pass with 2 errors
265mhz 3-4-4-8 memtest86 bandwidth = 3031MB/s test#5 ran for 31 passes, test #6 frooze after 2nd pass
263mhz 3-4-4-8 memtest86 bandwidth = 3008MB/s failed test#5, with 2 errors at 142nd passes
262mhz 3-4-4-8 memtest86 bandwidth = 2997MB/s failed test#5, with 4 errors at the 1140th pass ~20hrs
261mhz 3-4-4-8 memtest86 bandwidth = 2985MB/s stopped test#5 after 41 passes with no errors
260mhz 3-4-4-8 memtest86 bandwidth = 2974MB/s stopped test#5 after 228 passes with no errors, but freezes in 3dmark2003
255mhz 3-4-4-8 memtest86 bandwidth = 2917MB/s failed test#5 with 4 errors after 308 passes
250mhz 3-4-4-8 memtest86 bandwidth = 2860MB/s failed test#5 with 12 errors 1st time at pass 15, with 18 errors 2nd time at pass 19, while 3rd time ran error free for 32 passess befoe stopping [strange :confused: ]
245mhz 3-4-4-8 memtest86 bandwidth = 2802MB/s failed test#5 with 14 errors at 3rd pass [another strange result]
233mhz 3-4-4-8 memtest86 bandwidth = 2665MB/s
---------------------------------
==============================================
Unstable benchmarks of 4x 256MB ADATA PC4000 at 3-4-4-8 (both boths have memory performance modes disabled)
---------------------------------
Abit IC7
---------------------------------
233mhz 3-4-4-8
Sandra Buff = 5434/5392
Sandra Unbuff = 2887/2946
Aida Read = 5066
Aida Write = 1784
Sciencemark = 4557.19
Pcmark2002 CPU = 6946
Pcmark2002 Mem = 9620
Pcmark2002 Hdd = 1049
3dmark2001 = 11997
3dmark2003 = 3447
SuperPi 1m = 49s
SuperPi 16m = 19min 57s
Prime96 v23.6 benchmark = 71.972
Sandra CPU = 8646 / (2420/5167)
Sandra CPU multimedia = 12938/20416
250mhz 3-4-4-8
Sandra Buff = 5786/5795
Sandra Unbuff = 3273/3269
Aida Read = 5451
Aida Write = 2026
Sciencemark = 4861.63
Pcmark2002 CPU = 7409
Pcmark2002 Mem = 10575
Pcmark2002 Hdd = 1061
3dmark2001 = 12259
3dmark2003 = 3467
SuperPi 1m = 45s
SuperPi 16m = BSOD
Prime96 v23.6 benchmark = 66.948
Sandra CPU = 9320 / (2600/5551)
Sandra CPU multimedia = 13881/21027
255mhz 3-4-4-8
Sandra Buff = 5926/5833
Sandra Unbuff = 3189/3262
Aida Read = 5563
Aida Write = 2085
Sciencemark = -
Pcmark2002 CPU = -
Pcmark2002 Mem = -
Pcmark2002 Hdd = -
3dmark2001 = -
3dmark2003 = -
SuperPi 1m = BSOD
SuperPi 16m = -
Prime96 v23.6 benchmark = -
Sandra CPU = -
Sandra CPU multimedia = -
---------------------------------
Albatron PX865PE PRO2
---------------------------------
255mhz 3-4-4-8
Sandra Buff = 5347/5355
Sandra Unbuff = 3126/3151
Aida Read = 4672
Aida Write = 1963
Sciencemark = 4759.1
Pcmark2002 CPU = 7416
Pcmark2002 Mem = 9735
Pcmark2002 Hdd = 1034
3dmark2001 = 1st attempt frooze Car 1 test, 2nd attempt frooze Lobby 2 test
3dmark2003 = -
SuperPi 1m = 46s
SuperPi 16m = 18min 31s
Prime96 v23.6 benchmark = 66.185
Sandra CPU = 9290 / (2647/5885)
Sandra CPU multimedia = 14125/22357
260mhz 3-4-4-8
Sandra Buff = 5483/5467
Sandra Unbuff = 3127/3218
Aida Read = 4828
Aida Write = 1848
Sciencemark = 4877.94
Pcmark2002 CPU = 7749
Pcmark2002 Mem = 10283
Pcmark2002 Hdd = 1049
3dmark2001 = 13985
3dmark2003 = Frooze
SuperPi 1m = 45s
SuperPi 16m = error in 11th loop
Prime96 v23.6 benchmark = 64.894
Sandra CPU = 9619 / (2699/6028)
Sandra CPU multimedia = 14410/22686
270mhz 3-4-4-8
Sandra Buff = 5705/5701
Sandra Unbuff = 3362/3417
Aida Read = 5020
Aida Write = 2303
Sciencemark = 5172.55
Pcmark2002 CPU = -
Pcmark2002 Mem = -
Pcmark2002 Hdd = -
3dmark2001 = -
3dmark2003 = -
SuperPi 1m = BSOD
SuperPi 16m = -
Prime96 v23.6 benchmark = 62.486
Sandra CPU = -
Sandra CPU multimedia = -
tested on
Albatron PX865PE PRO2 @ 2.85v vdimm
Abit IC7 @ 2.8v vdimm (2.76v real idle vdimm)
==============================================
2.4C
MCX4000 rev1 + 120x35mm Sunon 120cfm fan + TT fan speed controller
128MB Gigacube 9600PRO at default
120GB WD 1200JB
Abit IC7 15 bios
vdimm = 2.8v bios (real 2.76v idle voltage)
vAGP = 1.6v
AGP/PCI: Fixed 66/33
NB Strap: 800FSB
HT enabled
Game Accelerator: Auto
Refresh Cycle Time: Auto
Read Delay Adjust: Auto
Read Delay: Auto
Command Per Clock: Auto
PCI Latency: 64
SATA: disabled
Firewire: disabled
==============================================
2.4C
MCX4000 rev2 + 92x38mm Sunon 120cfm fan + TT fan speed controller
128MB Gigacube 9600PRO VIVO at default
120GB WD 1200JB
Albatron PX865PE PRO2 1.02 beta bios
vdimm = 2.85v bios
vAGP = 1.6v
AGP/PCI: Fixed 66/33
Enhanced Performance disabled
Extreme timings disabled
OC Stability DDRAM disabled/blanked out
HT enabled
Spread spectrum disabled
Promise IDE Raid disabled
SATA controller disabled
Voice Genie enabled
==============================================
notes:
1. Single and 2 stick max oc's have been so far around 274-277mhz at 2.85v on PX865PE PRO2 for memtest86 upto 75mins looping test #5 and #6. I'll be revisiting single and 2 stick ocs and memtest86 stability tests in detail later on.
2. Introducing 4 sticks sees max oc's drop considerably on both PX865PE PRO2 and IC7. I've memtest86 tested 4 sticks looping test #5 and #6 for upto 24+ hrs still haven't found the absolute fault free speed for 4 sticks as yet.
==============================================
Memtesting results (4x 256MB ADATA PC4000 @ 3-4-4-8):
So far I've tested memtest86 looping test #5 and #6 for upto 24+hrs and for 4x 256MB ADATA PC4000, oc'ing headroom is way down for 2.76-2.85v vdimm.
---------------------------------
Abit IC7:
Come to conclusion 4 sticks on IC7 won't work due to lack of vcore. Under load in windows vdimm at 2.8v in bios drops to as low as 2.75v. Managed to do some benchies at 233, 250 and 255mhz even though memtest86 errored out.
Memtesting 233mhz 3-4-4-8 at 2.8v (2.76v real idle voltage) failed, after 237mhz 3-4-4-8 failed memtest86 with 18 errors after the 191st pass of test #5 loop, approximately 3hrs 11+ mins into test.
IC7's bios limit doesn't allow for >255mhz 1:1 as the pc boot screen freezes at memory display part, or memtest86 disk doesn't boot and stays at boot screen.
255mhz 3-4-4-8 AUTO memtest86 bandwidth = 2920MB/s failed 1st pass with 4 errors
250mhz 3-4-4-8 AUTO memtest86 bandwidth = 2866MB/s failed 5th pass with 12 errors
248mhz 3-4-4-8 AUTO memtest86 bandwidth = 2838MB/s failed 11th pass with 4 errors
246mhz 3-4-4-8 AUTO memtest86 bandwidth = 2818MB/s failed 32nd pass with 16 errors
245mhz 3-4-4-8 AUTO memtest86 bandwidth = 2804MB/s failed 13th pass with 12 errors
243mhz 3-4-4-8 AUTO memtest86 bandwidth = 2784MB/s failed 10th pass with 12 errors
240mhz 3-4-4-8 AUTO memtest86 bandwidth = 2750MB/s failed 214th pass with 12 errors
237mhz 3-4-4-8 AUTO memtest86 bandwidth = 2715MB/s failed 191st pass with 18 errors
233mhz 3-4-4-8 AUTO memtest86 bandwidth = 2668MB/s failed 31st pass with 30 errors [weird more errors sooner]
200mhz 2.5-3-3-5 Turbo memtest86 bandwidth = 2539MB/s failed 1st pass with 1920 errors
---------------------------------
Albatron PX865PE PRO2:
Currently memtesting at 245mhz 3-4-4-8 at 2.85v, seems I'm getting strange memtest results with 4 sticks under 255mhz, errors are showing up quicker than when > 255mhz :confused: Managed to do some benchies at 255, 260mhz and 270mhz even though memtest86 errored out.
Memtesting 255mhz 3-4-4-8 at 2.85v, after 260mhz 3-4-4-8 frooze in 3dmark2003 and failed in SuperPi 16m test. Only memtest looped test #5 for 228 passes error free before stopping it myself. Probably, if i fan it for 24hrs it would of shown errors at 260mhz 3-4-4-8.
PX865PE PRO2's 2.85v vdimm helped alot more in holding some decent speed for 4x 256MB ADATA PC4000 sticks of memory compared to IC7. Unfortunately, PX865PE PRO2 has a max vcore of 1.6v (undervolts to 1.52v), but for 4 sticks the oc'ing headroom is limited to well under what 2.4C's max oc is for 1.52-1.6v vcore.
275mhz 3-4-4-8 memtest86 bandwidth = 3145MB/s failed memtest86 immediately
270mhz 3-4-4-8 memtest86 bandwidth = 3088MB/s failed memtest86 test#5, 1st pass with 10 errors
268mhz 3-4-4-8 memtest86 bandwidth = 3065MB/s failed test#5, 1st pass with 2 errors
267mhz 3-4-4-8 memtest86 bandwidth = 3054MB/s failed test#5, 1st pass with 2 errors
265mhz 3-4-4-8 memtest86 bandwidth = 3031MB/s test#5 ran for 31 passes, test #6 frooze after 2nd pass
263mhz 3-4-4-8 memtest86 bandwidth = 3008MB/s failed test#5, with 2 errors at 142nd passes
262mhz 3-4-4-8 memtest86 bandwidth = 2997MB/s failed test#5, with 4 errors at the 1140th pass ~20hrs
261mhz 3-4-4-8 memtest86 bandwidth = 2985MB/s stopped test#5 after 41 passes with no errors
260mhz 3-4-4-8 memtest86 bandwidth = 2974MB/s stopped test#5 after 228 passes with no errors, but freezes in 3dmark2003
255mhz 3-4-4-8 memtest86 bandwidth = 2917MB/s failed test#5 with 4 errors after 308 passes
250mhz 3-4-4-8 memtest86 bandwidth = 2860MB/s failed test#5 with 12 errors 1st time at pass 15, with 18 errors 2nd time at pass 19, while 3rd time ran error free for 32 passess befoe stopping [strange :confused: ]
245mhz 3-4-4-8 memtest86 bandwidth = 2802MB/s failed test#5 with 14 errors at 3rd pass [another strange result]
233mhz 3-4-4-8 memtest86 bandwidth = 2665MB/s
---------------------------------
==============================================
Unstable benchmarks of 4x 256MB ADATA PC4000 at 3-4-4-8 (both boths have memory performance modes disabled)
---------------------------------
Abit IC7
---------------------------------
233mhz 3-4-4-8
Sandra Buff = 5434/5392
Sandra Unbuff = 2887/2946
Aida Read = 5066
Aida Write = 1784
Sciencemark = 4557.19
Pcmark2002 CPU = 6946
Pcmark2002 Mem = 9620
Pcmark2002 Hdd = 1049
3dmark2001 = 11997
3dmark2003 = 3447
SuperPi 1m = 49s
SuperPi 16m = 19min 57s
Prime96 v23.6 benchmark = 71.972
Sandra CPU = 8646 / (2420/5167)
Sandra CPU multimedia = 12938/20416
250mhz 3-4-4-8
Sandra Buff = 5786/5795
Sandra Unbuff = 3273/3269
Aida Read = 5451
Aida Write = 2026
Sciencemark = 4861.63
Pcmark2002 CPU = 7409
Pcmark2002 Mem = 10575
Pcmark2002 Hdd = 1061
3dmark2001 = 12259
3dmark2003 = 3467
SuperPi 1m = 45s
SuperPi 16m = BSOD
Prime96 v23.6 benchmark = 66.948
Sandra CPU = 9320 / (2600/5551)
Sandra CPU multimedia = 13881/21027
255mhz 3-4-4-8
Sandra Buff = 5926/5833
Sandra Unbuff = 3189/3262
Aida Read = 5563
Aida Write = 2085
Sciencemark = -
Pcmark2002 CPU = -
Pcmark2002 Mem = -
Pcmark2002 Hdd = -
3dmark2001 = -
3dmark2003 = -
SuperPi 1m = BSOD
SuperPi 16m = -
Prime96 v23.6 benchmark = -
Sandra CPU = -
Sandra CPU multimedia = -
---------------------------------
Albatron PX865PE PRO2
---------------------------------
255mhz 3-4-4-8
Sandra Buff = 5347/5355
Sandra Unbuff = 3126/3151
Aida Read = 4672
Aida Write = 1963
Sciencemark = 4759.1
Pcmark2002 CPU = 7416
Pcmark2002 Mem = 9735
Pcmark2002 Hdd = 1034
3dmark2001 = 1st attempt frooze Car 1 test, 2nd attempt frooze Lobby 2 test
3dmark2003 = -
SuperPi 1m = 46s
SuperPi 16m = 18min 31s
Prime96 v23.6 benchmark = 66.185
Sandra CPU = 9290 / (2647/5885)
Sandra CPU multimedia = 14125/22357
260mhz 3-4-4-8
Sandra Buff = 5483/5467
Sandra Unbuff = 3127/3218
Aida Read = 4828
Aida Write = 1848
Sciencemark = 4877.94
Pcmark2002 CPU = 7749
Pcmark2002 Mem = 10283
Pcmark2002 Hdd = 1049
3dmark2001 = 13985
3dmark2003 = Frooze
SuperPi 1m = 45s
SuperPi 16m = error in 11th loop
Prime96 v23.6 benchmark = 64.894
Sandra CPU = 9619 / (2699/6028)
Sandra CPU multimedia = 14410/22686
270mhz 3-4-4-8
Sandra Buff = 5705/5701
Sandra Unbuff = 3362/3417
Aida Read = 5020
Aida Write = 2303
Sciencemark = 5172.55
Pcmark2002 CPU = -
Pcmark2002 Mem = -
Pcmark2002 Hdd = -
3dmark2001 = -
3dmark2003 = -
SuperPi 1m = BSOD
SuperPi 16m = -
Prime96 v23.6 benchmark = 62.486
Sandra CPU = -
Sandra CPU multimedia = -