• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Stability testing results coming in.....3000+ winchester on Neo2

D

Duvie

Elite Member
Many know I hit 2650 range by being ablt to post at about 1.44-1.45v and even saw I posted at 2.8ghz with 1.63v....

BUt that was biospost and now this is windows post with memtest, superpi, and prime95 testing...NO 3dmark yet as I haven't installed my 5800u...using my sons 5200 card right now so the other system is still intact....

All at HTT 3x
cas 2, 1t, 4, 8, 4 133mhz ram divder
3.3% over vid vcore setting

2250mhz (9x250fsb)

Superpi 1mb = 40sec
Superpi 32mb = 35min 33 sec

Sandra cpu = 10215 / 4763
Sandra MM = 21314 / 22992
Sandra Mem (333ddr) = 4857 / 4857

Prime tested for 2hours
memtested for 5 passes

___________________________________

2350mhz (9x261fsb)

Superpi 1mb = 38sec *** better then my P4 at 3.5ghz***
Superpi 32mb = 34min 26 sec

Sandra cpu = 10671 / 4992
Sandra MM = 22296 / 24065
Sandra Mem (348ddr) = 5057 / 5061

Prime tested for 2hours
memtested for 5 passes

____________________________________

2450mhz (9x272fsb)

Superpi 1mb = 37sec
Superpi 32mb = 33min 09 sec

Sandra cpu = 11069 / 5203 ***better then my P4 at 3.5ghz***
Sandra MM = 23192 / 25010
Sandra Mem (364ddr) = 5285 / 5293

Prime tested for 2hours
memtested for 5 passes

___________________________________

2550mhz (9x283fsb)

Superpi 1mb = 35sec
Superpi 32mb = 31min 48sec

Sandra cpu = 11562 / 5388
Sandra MM = 24119 / 26055
Sandra Mem (378ddr) = 5505 / 5504

Prime tested for 6hours
memtested for 5 passes
Povray v3.5 render test
___________________________________

2600mhz (9x289fsb)

Superpi 1mb = 35sec
Superpi 32mb = 31min 06sec

Sandra cpu = 11808 / 5508
Sandra MM = 24634 / 26570
Sandra Mem (384ddr) = 5636 / 5614

Prime 2 hr passed
povray....testing


The highest temp I have seen in prime is 58c from corecenter but reads 54c from sandra and is that same 4c that there is from idle bios to idle corecenter readings. Cpu fan seems to spin up at certain temps and most of the time during prime95 it stays quite low...i heard from a a reliable source the corecenter temps are reporting high....

2650mhz (9x294fsb)

Superpi 1mb = 34sec
Superpi 32mb = 30min 32sec
Sandra cpu = 12073 / 5620
Sandra MM = 25121 / 27140
Sandra Mem (392ddr) = 5752 / 5736

Prime 6 hr passed
povray....pass
memtest- passed 5 passes - previous bios test passed 11 or 12 I cant remember..I need to write this crap down more...

I went from 3.3% to 5.0% but can't figure why corecenter reports 1.55-1.56 which is a .08 increase. I think the bios is only reporting 1.53v if I remember from bios testing....I nees to try to get a lower jump. Maybe this can get me closer to 300fsb....


Testing 2700mhz now.....

2700mhz (9x300fsb)

Superpi 1mb = 33sec
Superpi 2mb = 1m 18sec
Superpi 32mb = 29m 58sec
Sandra cpu = 12267 / 5718
Sandra MM = 25599/ 27598
Sandra Mem (392ddr) = 5842 / 5829

Prime ....failed at 1.57-1.58v so I upped it to +10% and now I am getting 1.63v but has been running for 4hrs
povray....passed
memtest- passed 5 pass

highest temp with superpi seen so far is 53c in corecenter 49c w/ sandra/bios (1.58v-load 1.6v)
highest temp with povray seen so far is 54c in corecenter 50c w/ sandra (1.58v-load 1.6v)
highest temp is 4+ hours into prime95 at 1.63v to 58c.....

Odd thing they all agree on vcore this time at 1.57-1.58v idle and sometimes loading to 1.59v in superpi. Like I said this volts vcore gets stronger during 100% load...interesting, huh???

 
Alright ! Now what are your impressions of your first AMD rig ? (or the first in a while)

Edit Time to change your sig ??
 
cas 2, 1t, 4, 8, 4 133mhz ram divder = only 188 FSB with horrid timmings

Your low bandwidth/timmings is seriously impacting those scores. I assume you're just eliminating mem as a culprit while you find chips prime stabilty?
 
Yep those were just some safe timings to work with the 133divider which was to definitely take ram out of the equation as a possible hinderance..That will come dead last....

It has been jumping quickly as I kept jumping 11fsb a time and getting closer to 300fsb where I will be back to pc3200 levels....

What do you think I can ultimately do??? Cas 2, 1t then what for th eother 3 numbers???
 
yes something is off.. here is my oc

3500+ (130nm) on msi neo2 plat.
10multi*245Fsb
2450Mhz
cpu voltage: 1.525 & Cpu over vid 3.3%
4xhtt
memory at 166mhz 2.5-3-3-8-1t running at 204Mhz at 2.6volts

superpi 1mb=35seconds
3dmark03=12,008 (6800gt at 380/1000)
sandra cpu arith.=11,143/5,167
sandra mem bandwidth=6,033mb/s & 5,995mb/s

I can hit 2.6 with it, but I am using a zalman cnps7000a-alcu for cooling and it gets into low 60s on full load at 2.6 that I don't like 🙂 right now i hit 54-55 on full load. this is according to msi core centre of course. MBM gives them of 49 on full load and sandra agrees with mbm5. Not sure which is right. oh well.
 
NOthing is off....I am running 384ddr versus your 408ddr.....Iused the other settings cause I heard they were safe...I am sure I probably can do cas 2, 1t, 3,3,8 when it is all down...Maybe I will be doing this at 2700mhz....

2600mhz passed 2 hours of prime.....superpi 32mb....now doing a povray rendering....


Same as every well else in the end cpu clock will far out trump memory speed...the timings are a tweak but I can tell you I will go balls to the wall and hold 133mhz.....
 
Does anybody know how I can post my screenshots to my comcast 3mb of space from my cable modem???

or

Let me email the pics to you and you post them for me!!!

I have a cpu_z screen pic at 2601mhz with stepping info, vcore, etc....
 
I don't know what that ram is Duvie other than geil 3500 you mentioned in PM. I mean what IC's. AMD 64 likes Micron, C Micron G, and Samsung TCCD. These all overclcok better than even on Intel boards according to AT. They don't like BH5, inferion, and hynixB so much. Ultimatly below 250Mhz Micron G aka Ballistix PC3200 $129 right now at newegg is the best. 2-3-3 1 T no problem. If you want to run 1:1 the TCCD is your only real option.

Anyway, If you get off stock HSF and into copper I see no problem 247 prime stable @ 2600Mhz @1.55V based on your results so far. Which is 288 FSB. Id hope you can run 166 mem setting 240Mhz(480DDR). 🙂Either way Tras needs to be 10 (theres about 300 MB/s read and 100MB/s write difference between 8 and 10 in my testing at much lower chip speed, 2100mhz) And of course, Cas should be 2.5 or less and command rate needs 1T. Other than those somewhat manditory settings I guess it depends of your mem can handle 240mhz...... Guess no. Hope for 2.5-3-4-3-10 1T @240 which will give excellent bandwith ~6200MB/s
 
Yeah Mark and I had them up to 460mhz cas 2 the other evening and took it into 480mhz with cas 2.5 before I dropped the divider cause I was jumping at large leaps at the time posting into bios.....I never used more then 2.7v or auto


Unless I cansee the real tangible performance increase beyond sandra's unrealistic world I will try to get to 2700mhz stable with best cas 2 timings and dual channel pc3200....I will test at 2600 with 480ddr but likely with cas 2.5 to see if the performance is better cpu or memory higher....I think I know from experience which one will win in most of the apps I run....
 
TTT....

I may need to jump to bios 1.41mod to see if vcore reporting in corecenter will start matching bios a bit more...Bios increases about what I though it should with slight undervolting, but coercenter jumped to way to much IMO from going from 3.3% to 5.0%...It should have been like .03v from the early 3.3.% vcore.
 
Originally posted by: Duvie
TTT....

2650mhz (9x294fsb)

Superpi 1mb = 34sec
Superpi 32mb = 30min 32sec
Sandra cpu = 12073 / 5620
Sandra MM = 25121 / 27140
Sandra Mem (392ddr) = 5752 / 5736

Prime 6 hr passed
povray....testing
Click to expand...

Wow nice chip!
 
TTT...

Think I will take it back down to 2664 at 1.57v and tweak the memory a bit....I not sure I want 1.6+ vcore even if I could keep temps in low 50's to high 40's load....


definitely cpu was holding back at 2700mhz with 1.57-1.58v...Cause current ram settings which I held are still running fine and below their rated speed.... All te rails are tight and if temps are to believed I never saw anything past 52-56c load and that should be fine for testing....
 
Testing a bit of ram choices...

Holding 2600mhz or 289fsb...

Tried 166divider with cas 2...no boot

Tried cas 2.5 leave 1t and the rest the same....

It booted...

Sandra mem = 6985/6871

errored in super1b wit an Error Not exact round....This was stable at 133divider but does vcore help with theis as well...I have 2.8vdimm being use now....

I don't think this ram may be able to do cas 2 to about 450 and 480 on cas 2.5 and who knows form there...

Edit: I wonder if the memory controller being on chip that the vcore has as much to do with this as anything else....What do you think?? I say this cause memtest is about 1 pass done and no errors at those settings....
 
Duvie,

I really enjoy your posts. Just finished reading about
your stability testing results coming in.....3000+
winchester on Neo2.

Could you do me a favor and list the specs on that
rig?

What type of RAM?
HD?

Thanks

Jack
 
So do I.

I'd like to see some benchmarks compared to his P4 also HT/multitasking which he did extensive testing on P4 already... for comparison.
 
Soon...soon...

Busy working this week...Going to Vegas this weekend, and rebuilt my son's rig this weekend...
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top