Bud Of Mine Is Having Fun...UT 250GB...

[AsAs1n8eR]

Darn I messed up I was in a hurry and forgot to get to this page of the forum(I was cold and hurried to close the window ) :But here it is with CPU-Z and the main page of the DFI-Street:

[CPDMF]

Good enough for me. Good job.

[learners permit]

GM these cores see no real advantage above 1.65 Vcore they simply run outta steam there. Search the XS forums there is a long thread by early adopters of the Newark core and it's quite informative.

[TheGeekMistress]

ok, thanx guys, and LP for the XS forum info

 

so, it seems like i'll be running between 1.6~1.7vcore with me new chip. it'll be under Asetek Waterchill so the temps should easily be below 50C load. i keep my apartment temp about 70F and i have great airflow thru my CM Stacker.

 

i'm thinking i could possibly sustain 3.1~3.3GHz on the core. and maybe, just maybe, if my luck with getting a prime piece of silicon, i could run up to 3.5GHz, but that's a long shot.

 

TGM

[learners permit]

My personal experience and everything I've read indicates to me that these Newarks have a sensitive memory controller. This core has NEVER seen voltage beyond 1.55v&104% and that was only for a very short duration. I'm personally not out to break records or cpu's but I manage to get good clocks with finesse rather than brute force. It certainly isn't the shortest path to good speed but it ensures longeivity and I learn a lot in the process. The temperature threshold for clocking high with low volts on this core is about 105deg. Farenheit. I'm hoping to get to 3.2ghz at this voltage but only time and effort will tell.

[TheGeekMistress]

the way i go about overclocking is i don't touch the voltage.

 

first i overclock and do a stability test by 5mhz on the HT bus at a time.

once the system becomes flakey, i look at the ram timings; if not the prob. i increase the vcore by .025 then test for stability again. i keep doing this until no more voltage will help. then i go about tweaking the other setting to make the highest clock speed increase little by little; for example messing with the mem tmings, or changing the multi and HT bus but keeping the same cpu frequency.

 

that's a gist of how i go about overclocking. i'm a very methodical person and i have always been that way. it usually seems to allow me a better overclock than most my peers as they just want quick answeres and to clock it up like the experienced people do that take the teadious testing time to get where they are.

 

TGM

[vector7]

ok, thanx guys, and LP for the XS forum info

 

so, it seems like i'll be running between 1.6~1.7vcore with me new chip. it'll be under Asetek Waterchill so the temps should easily be below 50C load. i keep my apartment temp about 70F and i have great airflow thru my CM Stacker.

 

i'm thinking i could possibly sustain 3.1~3.3GHz on the core. and maybe, just maybe, if my luck with getting a prime piece of silicon, i could run up to 3.5GHz, but that's a long shot.

 

TGM

 

My personal experience and everything I've read indicates to me that these Newarks have a sensitive memory controller. This core has NEVER seen voltage beyond 1.55v&104% and that was only for a very short duration. I'm personally not out to break records or cpu's but I manage to get good clocks with finesse rather than brute force. It certainly isn't the shortest path to good speed but it ensures longeivity and I learn a lot in the process. The temperature threshold for clocking high with low volts on this core is about 105deg. Farenheit. I'm hoping to get to 3.2ghz at this voltage but only time and effort will tell.

 

 

In XS I posted a couple of different CPU-Z shots of what I was able reach. However didn’t get the chance to share the details of what it took to get there.

 

I ran these benchmarks in two different environments. I eluded too the final one in my last post in Ozzimark’s 3700 Newark thread. One was with a 75F room temperature and the other with the window open. On the SLK 984U I found about a 20F spread from idle to load in both environments.

 

To effectively run past 3200MHz you need nearly 1.7v. I tried even more voltage even at 80F it made no OCing difference. If anything it made longer benchmarks unstable. However phase cooling could prove to be different.

 

I ran the following benchmarks just for stability purposes only and found that each benchmark would raise your idle temps as follows on my HSF. This explains where the CPU is headed in relation to your idle temperatures and why you’re able to complete them at different speeds.

CPU-Z=+3-5F

SPI (1M) =+15-20F

3Dmark2K1SE=+20F

 

Here's the highest MHz I could complete benchmarks without failing.

 

At room temperature 75F:

1.550v +104=1.58v

102F idle

CPU-Z=+3-5F=3212MHz

 

At room temperature 75F:

1.550v +114=1.68v

105F idle

CPU-Z=+3-5F=3293MHz

SPI (1M) =+15-20F=3185MHz*

3Dmark2K1SE=+20F=3133MHz*

I had to lower the core to 1.550v to get longer benchmarks stable enough to complete because the CPU was loading past 50C.

 

Window open ??F:

1.550v +114=1.68v

80F idle

CPU-Z=+3-5F=3388MHz

SPI (1M) =+15-20F=3289MHz

3Dmark2K1SE=+20F=3212MHz*

*Only ran it here because 1:1 stops for me at 247FSB…I know I should have run it at 9/10.

 

You can see at 100F idle you can find kill shots in the mid to upper 3200’s but you aren’t going to live there. But if you drop your load temps to 100F you could live in the 3200-3300MHz territory.

 

At 80F idle I found kill shots at nearly 3400Mhz but if you could drop load temps to 80F you probably could live in the 3300-3400MHz territory. This CPU with phase cooling should easily make 3500MHz if not more.

v7

[TheGeekMistress]

V7,

thanx for all that good data, it's very appreciated

 

sounds like water and a 220 pelt would make 3.3~3.4ghz an everyday reality.

 

water alone sounds like a max of 3.0~3.15ghz+ ..3.2ghz everyday would be nice on water.

 

once i get my system up and running i'll post results as well. but i've got to do some waterblock modding, buy the 4000 chip at the 1st of next month, and get my water system primed and running before i can even get it ready for a freash OS install. then i can finally get to figuring out how to tweak it out right since it's a brand new chipset/mobo for me.

 

i'll get there tho, i always do

 

TGM

[TheGeekMistress]

hey guys,

don't know if anyone's found this chart yet, but i screen cap'd it right off AMD's website HERE. It shows the 4000 mobile is only a 62watt core and was made in Dresden FAB plant and some other interresting stuff about internal bandwidth and such.

 

 

TGM

[learners permit]

Preliminary results are updated in my sig. This core doesn't scale with voltage as well as the 3700 does but the core speed is starting to get there albeit at a little higher volts than I had hoped.

[smolt]

Preliminary results are updated in my sig. This core doesn't scale with voltage as well as the 3700 does but the core speed is starting to get there albeit at a little higher volts than I had hoped.

3133 MHz hmmm and what volts is that ?

[TheGeekMistress]

LP,

seems to be yours is responding just as V7 had presummed. you're doing this on air right? my thoughts are 3.2GHz daily on water.

 

temps will prolly calm down by about 2~3C once the HS grease settles.

 

TGM

 

At room temperature 75F:

1.550v +114=1.68v

105F idle

CPU-Z=+3-5F=3293MHz

SPI (1M) =+15-20F=3185MHz*

3Dmark2K1SE=+20F=3133MHz*

V7,

thanx for all that good data, it's very appreciated

 

sounds like water and a 220 pelt would make 3.3~3.4ghz an everyday reality.

 

water alone sounds like a max of 3.0~3.15ghz+ ..3.2ghz everyday would be nice on water................>