Printable View
Venice hates Hynix - even with the memory on a divider the PC4000 was under DDR500 and the PC would not do anything but cycle a post. Worked fine with Winchester. Just gave it a go to see if I could do anymore with my Venice. Someone mentioned the DFI Ultra-D not liking hynix, thats not correct, venice just doesnt seem to.
Dean
FINALLY up and runnin, not totally loaded yet as I started playin with the ram. I have got some very interesting results so far. 1st for my setup...
Epox 9NDA3J rev 2.1 (week 4/05), heatsinks installed on all 8 mosfets, 6-cpu, 2-ram. now runnin 5401 bios. Alpha PAL 8150 with TT smartfan at (manual) 4000rpm. TT 480watt W0010 PS with PFC.
3200 venice (retail) LBBLE 0515GPBW
512*2 GSkill 2-3-3-6 econo pc3200's slot 1&2
2*WD800JB 80gig SE. IDE drives on primary M/S, not raided at this time.
Sony 16x DVD, NEC-2510 8x on secondary M/S
9800pro, ATI TV-wonder slot-6, SB-Live 5.1 slot-5, 3-com 10/100 NIC slot-3.
System booted at 2Ghz with 4b08(12-8-04) bios from factory. Vcore default was 1.40v with this bios. Disabled all unecessary on-board devices (NIC, sound, serial ports). All timing/voltages at auto for now.
Re-Booted from DOS floppy and ran update.bat for flash to 5401. Vcore now default is 1.35v...
Loaded WinXP-SP1 promptly without issue, Nvidia 5.10 drivers GART-IDE-SmBus ONLY, DX-9c ATI cat 5.4's, xp-sp1 defaults for others at this time.
UDSM 3.1 shows 2.708 VDimm as (auto) in bios. Set to 2.6 bios (2.608 UDSM) as it appears that this board does not undervolt RAM... :)
Ran memtest86 test5 only version overnight at 2-3-3-6-1t (2.6v). Looked schpanky this morning. Now runnin prime95 torture before I came to work,,, at 100% cpu max temp 37C... :D
Not overclockin cpu yet cuz I wanna lower cpu multi to 8X, htt to 3x and ramp up clock to find max ram settings tonight at 2.5(3)-4-4-8(10)-1t(2t) which will take some time.
So far I'm very pleased with the results and lookin forward to :slapass: this cpu. It appears from AMD's data sheet that the 3200 venice's have either a 1.35 or 1.40 default vcore as they define it as (variable). I'm hope'n lower vcore = lower heat = higher head-room??? time will tell.
Quote:
Originally Posted by bonzer
I don't agree with your last sentence, if you want lower vcore, how are you going to do make it work harder? You want to have lower heat go for watercooling or get good HSF or even phase change. :)
What I'm gettin at is that the venices can be burned at the factory for either 1.35 or 1.40, my theory would indictate that the cpu's passing all the factory QC tests at the lower default vcore voltage are likely to have a higher margin of overhead at increased vcore voltages.Quote:
Originally Posted by webbyboy
In other words this cpu required the lower voltage to pass AMD QC before it was shipped. Hence a very good possibility it will require less (overvoltage) to obtain the same overclocks as the 1.40 defaulted cores.... And I still gotta .05 vcore step to play with.... :toast:
like I said time will tell................... I will keep you posted
Is 1.35v printed on the IHS ?
The only 1.35v I saw as a 4000+ San Diego
Quote:
Originally Posted by ReelMonza
No its not however I believe it may be reflected in the last 4 letters???
3200 LBBLE 0515GPBW
check out this link for the default vcore(s) versus steppin
http://www.xbitlabs.com/articles/cpu...-venice_3.html
Quote:
Originally Posted by bonzer
Sadly the venice chips are just like the Winchesters I've had, I've had 9 of them looking for a good overclocker. I picked up a Venice to play with. They only go so far on stock volts, once you start upping the vcore quickly run out of 'headroom'. IE: one of my 9 winnies would need 1.488v to hit 292 HT, 1.53v to run 294, 1.58v to do 295 and 1.63v to run 297 HT. You hit that wall and quickly need more vcore to move 5 mhz on the HT which accounts to about 40 MHZ overall CPU speed. Is 40 MHZ really worth .1 volt? is it worth the added heat?
The Venice are the same way, regardless of burn in at the factory etc, the headroom theory isn't true. I've got 2 winnies here that can run @ 300 x 9 one needs 1.65v and the other needs 1.53, the Venice could not run stable past 2.65 Mhz regardless of vcore. They are just as much hit and miss regardless of what the stock settings show. Mine is a 3200 that boots and runs stock at 1.37 CPU-Z and bios but it didnt mean I had anymore headroom.
Dean
Can we make some early conlusion which 3000/3200+ Venice Stepping and/or CPU Assembly Date does overclock best or tends to overclock best ?
Anything with PAW seems to be good, like the Winchesters. Anything other than that seems to be hit or miss, mine is LBBLE EPEW, anything with PEW seems to live up to that expectation and they stink (PEW). :slapass:Quote:
Originally Posted by TEDY
I already forgot my stepping. I run it @ 1.55v 2600mhz. Probably a PEW, I'll check when I rip down my rig tomorrow.
So if I understand you correctly you are basing your conclusions on the fact that you have run 9 winchesters and 1 venice. Sadly, your total lack of experience with more than 1 the venice core hardly justifies your means..... or qualifies you as the venice authority.. :slapass:Quote:
Originally Posted by dnottis
I feel sorry you got biffed 9 times on your winnies and apparrantly 1 more time with your venice.. Remind me to never go shopping or gambling with you.... :clap:
The last four letters have nothing to do with voltage.Quote:
Originally Posted by bonzer
If it's really 1.35v it should be printed on the IHS.
Look:
http://akiba.ascii24.com/akiba/news/...ages770989.jpg
Dean- you of all people should know not to make generalizations like that, especially this early on ;) How many Venices have you tried vs. how many winnies? I guarantee you'll find a Venice suitable for your needs in fewer tries than it took to find your winnie :p:Quote:
Originally Posted by dnottis
BTW, both my venice 3500+ and 3800+ samples defaulted to 1.35 and didn't have that stamp. @ bonzer -so what's the default vcore on your 3200? You said it first defaulted to 1.4, then 1.35?
Quote:
Originally Posted by ReelMonza
craaaappppp, I didnt even notice..... :slap: , I was much to preoccupied with the date/step codes before mounted my sink
Don't place too much emphasis on that - read my above post ;)Quote:
Originally Posted by bonzer
default vcore was 1.4 with the early 12-8-04 bios installed from epox.Quote:
Originally Posted by s7e9h3n
default vcore was 1.35 after I flashed to the latest Rev E. compat bios (5401)
my 3500 oem came 1 day early!!
ADA3500DAA4BP
LBBLE 0515FPBW
Z634434D5xxxx
first time i have seen the FPBW.
so after average 3200 venice from monarch(2700+ - see earlier posts and venice oc thread) i am trying one from newegg. will post on oc thread in a little while. fingers crossed.
very truly yours,
politenessman
Found an interesting post by hummer010 (moderator) at littleblackdog.com which ID's and speculates ALL 3 lines of an A64 cpu part number. Check out the 2nd quote in the initial post:
http://www.littleblackdog.com/viewto...asc&highlight=
Quote:
Originally Posted by s7e9h3n
I agree - but with the Winnies I gave up and bought two from people that had them doing 2.7 Ghz, my luck never got any better ;)
I've tried 1 Venice but after the Winnie fisco I'm done, I'll sit tight at 2.7g @ 1.53v. Unless every chip starts to deliver consistent overclocks I'll wait this round out.
Dean
Quote:
So if I understand you correctly you are basing your conclusions on the fact that you have run 9 winchesters and 1 venice. Sadly, your total lack of experience with more than 1 the venice core hardly justifies your means..... or qualifies you as the venice authority..
I feel sorry you got biffed 9 times on your winnies and apparrantly 1 more time with your venice.. Remind me to never go shopping or gambling with you....
I'm not saying Im an authority on Venice because I had one. But I definitely could be with the Winchester and cmon, honestly do you really think Venice is all that different that Winchester? Venice is winchester with a few revisions added to it, it's not new technology. Also, check the forums here there are as many people with max 2.5ish overclocks as there are with 2.8 Ghz overclocks - sounds like it's just as hit or miss as the Winchesters were. :nono:
Yeah, and it's a good thing I'm not a gambling man ;)
Dean
Quote:
Originally Posted by bonzer
Interesting link, thanks :D
I got a CBBLE EPEW from newegg :(
We'll see how it overclocks after i eat.
Quote:
Originally Posted by vapb400
:slap: Don't eat - OVERCLOCK it - this is XS :slapass:
lol I had a 516 EPEW too from Newegg - curious to see your results.
Dean
Maybe I missed it but out of curiosity what stepping are the retail 3200s that monarch is shipping?
Quote:
Originally Posted by mongoled
Turns out it wasn't a "0502 GPAW" I ended up with a CBBLE 0512DPBW and have removed the shim, and it is currently priming at 2900Mhz @1.7v. I will post screenies later
Damn, my Venice crashes on Prime at 2.5 ghz @ 1.45V
Too bad I forgot to check stepping before I mounted :(