Core i7 Overclocking by Batch #

[Sam_oslo]

[IMG]..[/IMG]

I dont think theres any guarantee you'll get a good chip just because its a xeon, but there where some very good week 45-49 xeons. Then again there where many good 920's from the same weeks, but maybe not as consistent as the xeons.

If you have a decent 920 I wouldnt buy a new W3520, unless its either a confirmed good batch (new) or confirmed good clocker (used).

PS: 16 min of prime does not meen its gone run prime at this vcore for hours

Good one. A few more minutes and it would pass the kill-zone.
Is this your good old #3848A352 or a new one? In case what's the tricks for such a low Vcore?

[laragirl83]

3924A346

19*200MHz @ 1.200V LLC ON, p95, LinX stable and all.

[Sniper]

Good one. A few more minutes and it would pass the kill-zone.
Is this your good old #3848A352 or a new one? In case what's the tricks for such a low Vcore?

Its the same 3848A352.
This run was with 6 mem mp 7-7-7-20, requires less vcore than 8 mp. Also I'v compensatet low vcore with slightly higher then needed vtt (1.3).
And by switching from air to water gives lower temp, requiring less vcore. A win-win situation

[madmax999]

batch: 3849A837

24/7 OC
200*20 - 4GHz HT on, 1.26V vcore, QPI/VTT and PLL @ default values
200*21 - 4.2GHz HT on, 1.3V vcore, QPI/VTT and PLL @ default values

highest 3DMark 06 frequency
4.4GHz HT on, 1.43V vcore, 1.25V QPI/VTT, 1.9V PLL

highest OC with HT ON:


highest OC with HT ON:


max BCLK (for now, with 1.35V QPI/VTT and 2.0V PLL):


All done with TRUE and 2x120mm Antec fans

[Sam_oslo]

Its the same 3848A352.
This run was with 6 mem mp 7-7-7-20, requires less vcore than 8 mp. Also I'v compensatet low vcore with slightly higher then needed vtt (1.3).
And by switching from air to water gives lower temp, requiring less vcore. A win-win situation

[IMG]..[/IMG]

I see, interesting findings. Actually after playing with my CPU for a wile, I was expecting it to get stable @4.3GHZ(205x21) 1.24v. But even if it runs pretty much everything I use daily (and feels stable too) with 1.23v, but I couldn't get it Prime-stable before hitting 1.28v Linx is much easier than Prime on my CPU too (but maybe it got something to do with my MB, I guess).

Uncore:QPI 1:1 both at x18 @3.7GHZ is running stable with VVT=1.21v(BIOS set) and therefore I didn't try the idea of trying higher VVT as much, yet. My MB has a tendency to over-volt the VVT too, and 1.21v BIOS set will report 1.28v in BIOS, and that's another reason for keeping the VVT at low volt.

But your findings are interesting, and I may soon get time to give a shot on higher VVT to see if it will Prime with lower Vcore with 12GB RAM at x6.

What's your Uncore, and what Vdimm are you using?

[Sniper]

14x uncore and 1.64 vdimm. 30 min of LinX is about as hard as 2-3 hours of prime on this cpu/mb. Never got my DFI to play nice with prime (LinX was np), so I'v sold it.

[imo]

I dont think theres any guarantee you'll get a good chip just because its a xeon, but there where some very good week 45-49 xeons. Then again there where many good 920's from the same weeks, but maybe not as consistent as the xeons.

If you have a decent 920 I wouldnt buy a new W3520, unless its either a confirmed good batch (new) or confirmed good clocker (used).

PS: 16 min of prime does not meen its gone run prime at this vcore for hours

Very nice looks like xeon good batches are the dominators of SKT1366 overclocking ... now the question is where you can get one of those apart from you (if)... tried to match that vcore and this is what i got from the i7 920 before the lovely blue screen ... looks like its also thirsty for water

[imo]

what batch is that ?

3845B027

[imo]

Got few others too... but my hands are now eaching for a xeon 3848A352

[imo]

Xeon w3520 and i7 D0 both are very dependent on the weak of the production ( CPU-batch), but maybe you can have a better luck to find a good Xeon batch. My w3520 (#3848A352) is not the best batch, but is a decent/good one.

Yes very tempting...

[Sniper]

You have a very good chip there imo, not that many can match yours. Even if you could find a brand new 3848A352 there is no quarantee that it is as good as mine - or it could be even better. I had two of these 3848A352, and the other one was aproximatly 50-70 mhz behind this one.
In a couple of months we will have i9, and maybe 32nm i7?, then the race is on again and everyone gets dealt new chips.

[imo]

You have a very good chip there imo, not that many can match yours. Even if you could find a brand new 3848A352 there is no quarantee that it is as good as mine - or it could be even better. I had two of these 3848A352, and the other one was aproximatly 50-70 mhz behind this one.
In a couple of months we will have i9, and maybe 32nm i7?, then the race is on again and everyone gets dealt new chips.

lol I think your absolutely right my friend i should be happy with this 24/7 settings beside its not the best setup im running on air !
Im not a pro overclocker either... ppl use chilled water phace change dominator gt classified days after days of tweaking and then comesup with little less vcore or a number around 5Ghz...



I collected 5 great batches so far within the past 6 months.... sold 2 of my other batches to local mates ...
last one sold today 3845B (offer was too good to resist) and only two 3849B remaining one for me and one for...perhaps you...
They all are known to hit big numbers with proper setup.

[Sam_oslo]

As far as I have manged to follow, 3845B027 is one of the best batches out there. But a good max 24/7 OC would depend on other components like MB, RAM, and even GPU too.

The PCIe or the GPU (I'm not sure which yet), seams to be making some problems when the CPU hits @4.3GHZ(205x21). I have been testing with several different settings, Uncore (14x, 15x, 18x), VVT from 1.21v to 1.32v, Vcore=1.23v to 1.26v, and x6 RAM, but it has always been giving BSOD 3 minutes into Prime (almost on spot every time).

It Primes easily @4.3GHZ for about 3 minutes with Vcore=1,23v and up, but then suddenly the GPU core clocks raises from 2D to 3D, and a BSOD follows right at the same time. It will only pass when the Vcore hits 1.28v.

I'm not sure yet, maybe it is the GPU (GTX 295), or maybe the PCIe ? or maybe even my 12GB "mixed" RAM?. But I'm still working on it and hoping if I find the problem then the CPU would run stable with less Vcore.

[Sniper]

Imo, pimping your cpu's here with 16 min prime runs on a mb known for throtling is kind of dubious if you ask me. Why dont you use the for sale section ?

[imo]

As far as I have manged to follow, 3845B027 is one of the best batches out there. But a good max 24/7 OC would depend on other components like MB, RAM, and even GPU too.

The PCIe or the GPU (I'm not sure which yet), seams to be making some problems when the CPU hits @4.3GHZ(205x21). I have been testing with several different settings, Uncore (14x, 15x, 18x), VVT from 1.21v to 1.32v, Vcore=1.23v to 1.26v, and x6 RAM, but it has always been giving BSOD 3 minutes into Prime (almost on spot every time).

It Primes easily @4.3GHZ for about 3 minutes with Vcore=1,23v and up, but then suddenly the GPU core clocks raises from 2D to 3D, and a BSOD follows right at the same time. It will only pass when the Vcore hits 1.28v.

I'm not sure yet, maybe it is the GPU (GTX 295), or maybe the PCIe ? or maybe even my 12GB "mixed" RAM?. But I'm still working on it and hoping if I find the problem then the CPU would run stable with less Vcore.

Yes had to let it go to save my friendship lol nevertheless 026 is almost in pair with 027.

GTX 295 is definately it reminds me of my 4870x2 holding back the oc...raised pcie a lot and ended up frying the hdd, mixed ram not an ideal oc setup, therefore your cpu has more potential...

[imo]

Imo, pimping your cpu's here with 16 min prime runs on a mb known for throtling is kind of dubious if you ask me. Why dont you use the for sale section ?

This thread is about batches isnt it?
16min prime / linx is not that i post all the time you can look my previous posts in different threads. In Core i7 920 rev D0 stepping 5 thread when you requested a all memory 20 run linx I did, took me nearly an hour (how could you forget) would want to see how your batch does with linx all test,

Yes but got the 0006 bios now for no throttling

Thought to let ppl know as Im getting reed of my batch collection, not that i expect a real sell from the forum but never the less doesnt mind either if any body looking hard but again all my post here are ABOUT BATCHES. and 16 min gives a pass on prime to take a snap as you did aswell as i dont know of any restrictios that you cant

Come to think of it yes I might...

[FtW]

Haven't seen my batch 3849a832 in here (bit of a long read though, could've missed it).
Here's some 20 pass linx results, been hooked on benching lately & haven't made time for a long prime run yet...

4.2 ghz, 1.248V idle 1.216 loaded


4.4 ghz, 1.360 idle 1.328 loaded

[nascasho]

Man, I' been shopping around for one of those i7's that'll hit 4.2Ghz on super low voltages that you guys get! They're hard as hell to find!

[Osterman]

I want to ask.
Does someone have a XEON W3502 with this batch: 3901A.
Because I have the oportunity to get one of those and would be nice to know if it is worthy.

Mi actual cpu a 920 D0 and the batch begins with 49, does 4.2 with 1.36875 vcore on BIOS with HT enable.

[Godmyster]

hey guys i just got myself a D0 its new, batch# 3918A478 anyone here have one of these?

[Aphellyon]

Hmmm... mine's 3917C045

[Godmyster]

i wanted some help on the configs for the UD5, my goal is 4.2 atm

[Fafeifa]

Bought a core i7 3911A237 off craigslist for $150, it does 4.2Ghz rock stable with vcore 1.25v under load qpi is 1.3v, haven't bothered to go higher until finish building my loop

[Osterman]

Wich one is your motherboard?
Because I also have one of does but... it need a lot more vcore :p

[Salamndar]

Hmmm... mine's 3917C045

I am not aware of a C batch, I only thought As and Bs !!