Do Micro Center CPU's OC as well as standard retail CPU's?

[Ferzerp]

Originally Posted by Acanthus

but a known good bin has a better chance for a good chip than a known bad bin.

That's never really been statisically proven. In fact, I was under the impression that it had entirely been dismissed anyway.

[Acanthus]

Originally Posted by Ferzerp

That's never really been statisically proven. In fact, I was under the impression that it had entirely been dismissed anyway.

I know back in the day there were websited dedicated to selling "better bins" for a slight markup.

I know for the Core 2 Quads and Core 2 Quad based Xeons the G0's were far far better than the B3's... and i know my SLACT Xeon did very well as did most others.

[Ferzerp]

Originally Posted by Acanthus

I know for the Core 2 Quads and Core 2 Quad based Xeons the G0's were far far better than the B3's...

Those are steppings and not batches.... The other part of your post is just anecdote.

[BrightCandle]

Back in the Athlon/Duron days (800Mhz - 1.4 Ghz) there were two plants making the chips. The Dresden plant produced chips that had a noticeably better overclocking potential than the others. They also had a slightly greenish hue around the die instead of a blue one, so you could if looking at chips determine where it came from. You were still subject to centre or outside wafer variance but statistically the Dresden chips were better by about 50Mhz (which on 800-1000 was a lot).

Nowadays I see there is a lot of batch tracking for the modern chips and nothing obvious is coming out of the data right now. I would be surprised if later batches didn't have better quality, Intel (and Global Foundries) are constantly updating the production line to reduce the number of dead chips and thus increasing the quality of the product on average.

For overclocking on the package it would be nice to have Intel testing the chip at various overclocked voltages and the clock speed it achieves. Even if all they did was go up to 4.0 Ghz that would give us some idea of the curve. The default vids with boost vids would likely give us enough of that curve to have an idea what it would look like. In addition I would really like to know where from the wafer the chip came, because chips from the centre of the wafer are often much better (at least that is what the experts tell me).

[Acanthus]

Originally Posted by Ferzerp

Those are steppings and not batches.... The other part of your post is just anecdote.

It is absolutely true that it is anecdotal evidence supporting it.

I have had a good experience with it. That is all I am saying.

There rumors that "C" batch Sandy Bridges do better than others. Even though they are all the same D2 stepping.

http://www.tomshardware.com/forum/26...g-overclocking

http://www.overclock.net/t/1179326/a...he-better-ones

Of course again it is anecdotal and there are a lot of variables like Motherboard, Cooling, Ambient Temps, Power Supply, etc.

[VirtualLarry]

Originally Posted by BrightCandle

For overclocking on the package it would be nice to have Intel testing the chip at various overclocked voltages and the clock speed it achieves. Even if all they did was go up to 4.0 Ghz that would give us some idea of the curve. The default vids with boost vids would likely give us enough of that curve to have an idea what it would look like. In addition I would really like to know where from the wafer the chip came, because chips from the centre of the wafer are often much better (at least that is what the experts tell me).

That gives me an idea. Why not read out the Vid of SB chips at the various multis, graph it, and give a rating as to how overclockable your CPU is.

Kind of like how GPU-Z gives you an ASIC quality score.

[ehume]

Hmm. This is interesting. On the one hand, I have IDC's insight that MC will have some returns salted in among the unused chips, with an unknown percent not-overclockable. OTOH, we have a recent 3770K that does 4800MHz on a 1.216v Vcore.

Then I know that the 3770k's are good until they need Volts, and then they get too hot to cool because there is TIM under the IHS, not solder. But to get a good-performing Haswell I'll have to wait seven or eight months.

Quite the decision.

[LagunaX]

Mine is 4.8ghz under 80c load on air NOT delidded (though soon...).

Roll the dice...

But read these 2 threads (Dec. 2012 onwards) to hopefully avoid the more mediocre numbers, some batches are truly crappy:

http://www.overclock.net/t/1348988/i...atch-number/70

http://www.xtremesystems.org/forums/...results/page22

[Firestorm007]

Originally Posted by IntelEnthusiast

We only sale to our Distributors (i.e. Ingram Micro, Tech Data, Synnex, etc.); who than in turn will sale to the reseller (i.e. Newegg.com, Microcenter, Amazon, etc.). So all the processors for all the resellers are all coming out of the same big stock that goes to the disti. So no there is no differences from one to the other.

Thanks, Christian. I always appreciate your feedback. Just wanted to let you know.

[Torn Mind]

Has anyone ever thought to ask themselves how can Microcenter employees (or Newegg or Amazon) know which boxes to choose or if they even have the chance to see the boxes? Perhaps some wafers manfactured at different times can certainly have crappier chips, but can they tell once it is released to the shipping channels. I doubt it.

[LagunaX]

The sales associate is usually happy to look for a certain batch if I ask for it since they work partially on commission at Microcenter also, or give me a choice of what is available, the boxes are in a back office. Fry's is usually more difficult as they are in a large cage.

XS and TPU member Dumo probably has a friend that works at Microcenter also.

[Torn Mind]

Originally Posted by LagunaX

The sales associate is usually happy to look for a certain batch if I ask for it since they work partially on commission at Microcenter also, or give me a choice of what is available, the boxes are in a back office. Fry's is usually more difficult as they are in a large cage.

XS and TPU member Dumo probably has a friend that works at Microcenter also.

I meant from the distributor Micro Center buys from.

But, very interesting info.

[LagunaX]

That's how I asked for and easily got a second batch 3231B415 3770k retail chip ( same as the above Prime 95 post) the other day...will test it over the weekend :-)

[dawp]

I know my MC i7 950 seems to be a decent chip, stable @ 4.2 with 1.4v.

[blastingcap]

Originally Posted by Idontcare

In theory there ought to be no difference between buying a retail chip from MC versus Newegg.

But there is one mechanism in play that has a bias towards MC and which can result in the common Joe having increased odds of getting a lemon - the return policy.

If I buy three 3770K's from Newegg, OC the hell out of them to find the most golden chip of the three, and return the other two to Newegg for a refund then I am out the restocking fee plus shipping for two cpus.

If I were to do the same thing by driving to a local Microcenter, then I am only out the restocking fee, but not the shipping and handling fee (just my gas and my time driving back to the store).

So it is reasonable to expect that to whatever extent this natural selection process is placing pressure on the outstanding inventories at hand in retailers versus etailers, filtering out the best of the best OC'ing chips faster at retailers than etailers because of the price differential that comes with the territory, then it does lead one to conclude that your odds of randomly getting a golden sample from a retailer are less than that of randomly getting one from an etailer.

That much it seems we are safe to conclude. What we can't deduce is just how much of an impact, statistically speaking, this bias introduces. Are your chances lessened by 0.1%? Or 10% Or 50%?

Unrelated to today's situation, but related tangentially, years ago I use to buy my CPUs from an etailer called atacom.com. Their differentiating value-add was that they would pre-test and bin CPUs for their OC potential - if you wanted "guaranteed" 550MHz OC from a 333MHz celeron then they were the people to buy from (at a price premium of course)...but you knew not to buy CPUs from their "stock inventory" because you knew they had already filtered through it and removed all the premium OC'ing chips from the population.

I think you can see how that anecdotal tale, a true one at that, relates to the hypothetical one I outlined at the beginning of this post.

I thought most shops have a no-return policy on CPUs? You can only send it back to Intel itself?

[lehtv]

Originally Posted by blastingcap

I thought most shops have a no-return policy on CPUs? You can only send it back to Intel itself?

Newegg has a replacement-only return policy on CPUs, so no refunds. If you were to sample their CPU collection for the best overclockers, you'd have to try them one at a time, or ebay the ones you don't keep

[dawp]

I'd be surprised if newegg would do that for more than a few times without having large purchases.

[Bill Brasky]

Originally Posted by lehtv

Newegg has a replacement-only return policy on CPUs, so no refunds. If you were to sample their CPU collection for the best overclockers, you'd have to try them one at a time, or ebay the ones you don't keep

Same with MC.

[Zap]

Originally Posted by LagunaX

The sales associate is usually happy to look for a certain batch if I ask for it since they work partially on commission at Microcenter also, or give me a choice of what is available, the boxes are in a back office.

That was true for me last time I was looking for something specific, when new Core i7-920 D0 just came out but plenty of old C0 still in the channel. The MC salesman let me pick based on the label.

Unfortunately my ASRock X58 Supercomputer board was on the original BIOS and won't POST with a D0.

Anyone want to buy an ASRock X58 Supercomputer board that doesn't POST with D0 chips? Essentially NIB and really cheap.

[Ferzerp]

Originally Posted by Zap

That was true for me last time I was looking for something specific, when new Core i7-920 D0 just came out but plenty of old C0 still in the channel. The MC salesman let me pick based on the label.

Unfortunately my ASRock X58 Supercomputer board was on the original BIOS and won't POST with a D0.

Anyone want to buy an ASRock X58 Supercomputer board that doesn't POST with D0 chips? Essentially NIB and really cheap.

Asus did a wonderful thing on their newer boards. You can update the bios via USB drive with nothing installed. All you need is power, and the USB stick formatted correctly with the BIOS named properly, and you can update it. I <3 that idea. It makes things like your situation a thing of the past.