How has Venice cpus changed your TCCD ram clock settings ?

[eva2000]

Question for Venice cpu using DFI NF4 users.

How has changing from winnies/newcastle/clawhammer cpus on DFI NF4 to venice cpus, changed the bios memory timing/settings used for your Samsung TCCD ram ?

Particularly with respect to Gskill PC4400LE TCCD ram, any changes to drive and drive data strength, TREF and vdimm used that differ when using Venice cpu ?

Which bios used, which dimm slots used ?

Care to share :D

[Eldonko]

For me settings for my TCCD and diego had to be totally different than with my previous Newcastle. I need more volts (3.0v vs. 2.8 with NC) to even boot into windows and I can get no where near the clocks I got with the old core. I have tried pretty much every BIOS and setting and my end result was losing about 40Mhz of OCing headroom.

When I tried booting at settings that I had saved for my previous CPU (was only 254Mhz @ 2.5-3-3) I screwed up my BIOS and had to clear CMOS and got no where near booting into windows.

From my experience, my advice is to basically start from scratch, use the settings you used before as a guide.

I liked 0414-3 the best for a BIOS, all of the dividers worked well as did the CPU multis.

From the posts I have read, many are having different experiences with their RAM when switching to Rev E, some good some bad.

[formyfaith]

For me settings for my TCCD and diego had to be totally different than with my previous Newcastle. I need more volts (3.0v vs. 2.8 with NC) to even boot into windows and I can get no where near the clocks I got with the old core. I have tried pretty much every BIOS and setting and my end result was losing about 40Mhz of OCing headroom.

When I tried booting at settings that I had saved for my previous CPU (was only 254Mhz @ 2.5-3-3) I screwed up my BIOS and had to clear CMOS and got no where near booting into windows.

From my experience, my advice is to basically start from scratch, use the settings you used before as a guide.

I liked 0414-3 the best for a BIOS, all of the dividers worked well as did the CPU multis.

From the posts I have read, many are having different experiences with their RAM when switching to Rev E, some good some bad.

Yup, a lot of my friends here using San Diego cores are having trouble

using TCCD rams compared to there Clawhammer, Winchester CPUs.

Through a bunch of tests and etc, so far it looks like DFI's bioses are not

completely suiting the new San Diego core.

Though not 100% completely sure about it.

[Aurhinius]

Got the same issues I cannot overclock anything like what I could with a winchester....

A few folks have had no issues but others are having nightmares like me. I have had both a venice core and a san diego, I have had corsair XMS4400 and now OCZ 4200EL Plat and niether will hit 250mhz no matter what timings or bioses I use.

Disable CPC and bank interleave and you get back to stability but at a huge cost to memory performance.

[[timko]]

Moving from my Winchester 3200 to Venice 3200 I was able to use the same DRAM settings to keep the HTT at 310Mhz 1:1 1t with my OCZ Rev.2. However, I lost the expected amount of approx 200 to 300points in Sandra Memory Bench.

Used the same...

• bios with both CPUs --> 510-2 (no that wasn't a type, I said five one zero two)
• two orange slots
• vdimm and all of the DRAM settings

I'm sure I could probably recover some of the lost Sandra mem points if I tweaked it a little but I'm not too bothered by it really as long as it's stable :)

[Eldonko]

Got the same issues I cannot overclock anything like what I could with a winchester....

A few folks have had no issues but others are having nightmares like me. I have had both a venice core and a san diego, I have had corsair XMS4400 and now OCZ 4200EL Plat and niether will hit 250mhz no matter what timings or bioses I use.

Disable CPC and bank interleave and you get back to stability but at a huge cost to memory performance.

Same issues for me, but I would rather run the memory at friggin stock than use 2T and disable bank enterleave. >:(

[Aurhinius]

I don't know what is up with it. I have tested the memory in memtest and it passes fine. I boot up windows and it won't do 32M superpi at all.

soon as the 2T and bank interleave off happens I can 32M superpi fine.

Dropped the FSB to 230 and could superpi fine. Anything over 230 it wouldn't superpi at all. Tried a memory divider FSB 255, actual memory 215 - no 32M superpi.

Figure my on die memory controller is crapping out at anything above 230. For a San diego that is pretty crap to be honest. Just seems wierd the venice responded pretty much the same.

Don't really want to buy a third chip to find out if I just had bad luck with the other two......

Dropped the multi of the CPU to keep it in stock range mhz but FSB of 255. Again no 32M superpi.....

Something strange going on here....

[XiN]

I haven't a Venice but a Winchester (a bad CBBID).

With bios 310 I had found a good stability at 300 mhz (memtest, Gold Memory normal test, prime, spi 32M etc...) with certain settings ... with bios 510-3/1 I have lost stability (many errors in GM and memtest) without changing dram settings ... so now I'm testing new settings and maybe I have found stability newly ...

In my opinion the problem is in the new oskar bioses and in the new ram tables ... looking at http://xtremesystems.org/forums/showthread.php?t=53290&page=1&pp=25 there are guys with Venice/San Diego who are stable with TCCD at 300 or more

[austinbmxnig]

i have a neo2 and my venice won't run 1:1 @ the g.skill le speeds (275 2.5-3-3-7), but the venice will do 2.8 stable on a divider

[burningrave101]

']Moving from my Winchester 3200 to Venice 3200 I was able to use the same DRAM settings to keep the HTT at 310Mhz 1:1 1t with my OCZ Rev.2. However, I lost the expected amount of approx 200 to 300points in Sandra Memory Bench.

Used the same...

• bios with both CPUs --> 510-2 (no that wasn't a type, I said five one zero two)
• two orange slots
• vdimm and all of the DRAM settings

I'm sure I could probably recover some of the lost Sandra mem points if I tweaked it a little but I'm not too bothered by it really as long as it's stable :)

What are you BIOS settings for your RAM? DRAM settings and voltage.

i have a neo2 and my venice won't run 1:1 @ the g.skill le speeds (275 2.5-3-3-7), but the venice will do 2.8 stable on a divider

I have a 3200+ Venice as well on a DFI Ultra-D with OCZ PC3200 Platinum Rev2 and mine will do 2.8Ghz but i have to use a RAM divider. Disabling CPC and bank interleave seem to really improve stability alot but it hurts performance. Maybe a new BIOS from DFI will help things, i dont know. I thought the memory controllers were suppost to be vastly improved with the new revision E chips? These things are CRAP!

[mad mikee]

I've had my LE @ 2.5-3-3-7 (1hr of OCCT / prime) and 2.5-4-3-7 (very stable - runs OCCT torture w/ 900MB setting for hours :D )
Between 100-200 MB/sec faster than my dream winnie (which could NEVER do 2.5-3-3-7 w/ any stability
Just need 1.62 volts to do it :slap: (Ordering another today :D )

[burningrave101]

I've had my LE @ 2.5-3-3-7 (1hr of OCCT / prime) and 2.5-4-3-7 (very stable - runs OCCT torture w/ 900MB setting for hours :D )
Between 100-200 MB/sec faster than my dream winnie (which could NEVER do 2.5-3-3-7 w/ any stability
Just need 1.62 volts to do it :slap: (Ordering another today :D )

Thats swell dude but what DRAM settings are you using in the BIOS?

Anyone that have a Venice/San Diego core and TCCD based memory post the DRAM settings your using.

[ccokeman]

I'm using the 510-1 and the max i can get stable is 258x11 fsb. I cannot get it to even run memtest above 272. I'm using G.skill 4400 le w/440 chips and am getting nowhere close to the fsb i could run with my winchester. This is what i have to use to get to where i am. 2.8v on the mem 1.62 on the vcore.

[theobub]

Thats quite odd.
I just got a Venice 3200+, the mobo is a DFI NF4 UT ultra D, G. Skill 2x256 TCCD F1... LE, and I have been doing fine. I can run benchs up to 9x313@1.60v (air cooled). But above that, and when say above I mean 314, the CPU needs more Vcore, but with around 1.776 it gets stable with up to 9x320, thou it gets a little bit too hot. This rig has been working for me so long. I will soon post some screenshot with my timings (thou I will be ashamed of the timings), but I hope it will help
:confused:

[AMDnewbie]

had a 3200 0517 vpmw here..
venice don't like tccd but like bh-5....
it allows bh-5 to run as tight as possible upwards but not tccd...

my pair can do 275Mhz @2.5-3-3-7 w winnie
but having problem crossing 265Mhz w same setting and 275Mhz now can only be done w 2.5-4-3-7...

[andyOCZ]

Strange, as I have found my Winnie CBBID to be quite similar to my Venice 3200 in all timings (except that the Aplha timings can be tightened a bit). This is on the DFI NF4 with OCZ TCC5 and TCCD. I have found that I can run 312 mhz with CPC off, but not above a 7 multi. With CPC on I can run 300mhz, but again only with a 7 multi. Any higher CPU overclock and i fail to post or having stabilty issues.

Then again my 2 CPU's are 90nm, and the Winchester is an 0503, so it's possible the CPU's are closer in design than we know.

[L@TrO]

I had a cbbid winny and now a lbble 517 venice. This venice is great clocker , but the mem controler is worse ( or less good :) ). Use the same vdimm , same tref ,but I have to put the drive data strength to 8ns ( 7ns before ) when I go for high freq ( 300+ mhz).

note: I'm using the 510-3 bios and slot 2&4

[burningrave101]

What Drive Strength and Data Drive Strength values are you guys using for your TCCD memory with a Venice chip?

[theobub]

What Drive Strength and Data Drive Strength values are you guys using for your TCCD memory with a Venice chip?

I am leaving them to auto in bios 510-3, those are my timings, hope it helps: