PDA

View Full Version : Z68 suspicious revision numbering?


Moshing
05-10-11, 05:24 PM
By now, almost anybody who's been watching the Z68 hype building around it's launch know what it's claim to fame over P67 is, mainly the SSD caching and use of the iGPU simultaneously with a discreet GPU for encoding purposes. For the record though, neither of these require much or even ANY dedicated chipset hardware.

Now on to the suspicious part. I have a strong feeling that Z68 will offer no performance benefit over P67 except in those 2 scenarios where the new features are being used. But what is up with that B3 revision showing up all over the place? Seems pretty straight forward to me. Z68 is the exact same silicon as P67. Not a transistor different. I'd go so far as to accuse Intel of pulling them from the same wafer in fact. Chipset revisions are a telling factor when comparing multiple models within the same family. Same silicon gets same revision, regardless of binning. You can bin them out for various models, but if the base specs call for the same bus speeds across the line, then they will all perform similarly unless specific performance enhancing features are turned off for a specific family.

In this case, I'm thinking the WHOLE difference between H67, P67 and Z68 consists of a couple cuts with a laser and a firmware revision to recognize which cuts make it which model. Cut a line here, poof, H67 and P67 come to market not recognizing SSD caching. Cut a different line, poof, there goes overclocking and multiplier adjustments, and H67 is born. Cut a different one instead, and poof, the iGPU will no longer work alongside a discreet GPU, and you have a P67.

Moral of the story? I believe new purchasers will absolutely want Z68 if they don't mind paying a premium AND have need of either of those 2 new features. If you DO mind paying a premium because you have zero use for SSD caching or encoding, I strongly believe P67 will remain the bang for buck king in the Intel stable even after Z68's launch. And will most likely remain the preference of gamers.

Viral
05-10-11, 07:30 PM
This whole B3 revision thing has been marketed like it was a feature or something. It's almost like slapping HD on some random device even! They fixed a problem, it's not the second coming! Just what it should have been from the start.

I'm thinking because everyone knows that B3 = one to get, they've decided to tag it onto the Z68 so it doesn't miss out on the free marketing?

Moshing
05-10-11, 09:00 PM
This whole B3 revision thing has been marketed like it was a feature or something. It's almost like slapping HD on some random device even! They fixed a problem, it's not the second coming! Just what it should have been from the start.

I'm thinking because everyone knows that B3 = one to get, they've decided to tag it onto the Z68 so it doesn't miss out on the free marketing?

I don't think it's that at all. Pretty sure my theory was closer if not dead on target. Intel doesn't just slap useless revisions on chipsets to sell them.

Edit: Just as a point of reference, buying a new mobo when Z68 comes out isn't outside of my monetary capability. I just don't see the point, is my whole point. Should I be proven wrong on this, have little doubt I'll be picking one up as soon as there's an Asus Sabertooth Z68. If Asus doesn't do a Sabertooth of Z68, which is entirely likely as rumor points towards Asus' next Sabertooth not coming till X79 and socket 2011, then I won't give up my killer mobo just for iGPU encoding. And already having my OS on an SSD, the new SSD caching feature is worthless to me as well.

frenchy2k1
05-11-11, 11:56 AM
biggest claim to fame is having both integrated GPU AND overclocking enabled.
Fact is that you are probably right (i could even verify it if you wanted).
This is a known practice called "binning". Similarly, an i5 2500 is the same as a i7 2600k, they just have been tested (and fuse locked) to different capabilities.

It is quite possible that the Z68 is a finally debugged P67/H67. If the yields were bad for having both overclocking and iGPU, they create 2 products with each one a different feature enabled. This allows them to sell "imperfect" dies. Maybe the early revisions of P67/H67 were supposed to have those features, may be not.

All semiconductor manufacturers do the same and sell as much of their production as possible (see the GTX465 from nvidia for example, which sold their worst GF100 chips).

Bman212121
05-11-11, 06:18 PM
The P67 probably uses the same chipset as Z68 but all P67 boards lack a DVI connector to actually use the integrated graphics. There was obviously some reason why they had 2 different board options prior to Z68 (Like lucid not working properly yet) so they made two platforms instead.

Moshing
05-12-11, 11:39 AM
biggest claim to fame is having both integrated GPU AND overclocking enabled.
Fact is that you are probably right (i could even verify it if you wanted).
This is a known practice called "binning". Similarly, an i5 2500 is the same as a i7 2600k, they just have been tested (and fuse locked) to different capabilities.

It is quite possible that the Z68 is a finally debugged P67/H67. If the yields were bad for having both overclocking and iGPU, they create 2 products with each one a different feature enabled. This allows them to sell "imperfect" dies. Maybe the early revisions of P67/H67 were supposed to have those features, may be not.

All semiconductor manufacturers do the same and sell as much of their production as possible (see the GTX465 from nvidia for example, which sold their worst GF100 chips).

THIS has to be the BEST explanation I've seen yet. Binning was my suspicion as well, but I didn't want to lead anybody to conclusions if they had a better one then me. My only conclusion I was comfortable making was that they were the same silicon, but your binning explanation is the most reasonable cause.

Moshing
05-12-11, 11:41 AM
The P67 probably uses the same chipset as Z68 but all P67 boards lack a DVI connector to actually use the integrated graphics. There was obviously some reason why they had 2 different board options prior to Z68 (Like lucid not working properly yet) so they made two platforms instead.

Yup, well said. I think there's more to it though then just that. I think they outright laser cut the chipset to disable features in the same way AMD and Intel uses laser cuts to shut off defective cores on CPUs to be sold as lower end parts.

jcniest5
05-15-11, 10:52 PM
I was thinking the same thing and voiced my opinion on a different forum. The idea behind the B2 and B3 for P67 is that the latter has fixed the potentially problem SATA ports while B2 doesn't. Now, if Z68 is different than both B2 and B3, then why do they still need to put the B3 with the naming scheme? To me, B3 is bogus...it shouldn't even be there if Z68 is newer, everyone knows it's different. To put B3 there just confuses everyone.

Moshing
05-15-11, 10:59 PM
I was thinking the same thing and voiced my opinion on a different forum. The idea behind the B2 and B3 for P67 is that the latter has fixed the potentially problem SATA ports while B2 doesn't. Now, if Z68 is different than both B2 and B3, then why do they still need to put the B3 with the naming scheme? To me, B3 is bogus...it shouldn't even be there if Z68 is newer, everyone knows it's different. To put B3 there just confuses everyone.

But it's not different, that's the point. So the B3 is actually accurate. Z68 is the same silicon as P67 and H67 and possibly even comes off the same wafers.