Unknown IBM chips

[Vegeta]

I can not find any information about these chips. What's this?

IBM 70P5381 СLGA-1247
IBM 06K7145/06N4267 СCGA-1247
IMB 18K0080/29K4314 СCGA-1678

[Vegeta]

IMB 18K0080/29K4314 СCGA-1678 similar to POWER3-II?

[isa-d]

1) it's a support chip
2) one way RS-64 III
3) Power 3 II

[Vegeta]

Thank you!
Advice please how to remove of compound traces from ceramic borders? For example like this Power 3 II.

[Vegeta]

IBM p/n 06K7145 is present on the scan in this GB topic http://www.cpu-world.com/forum/viewtopic.php?t=20161&highlight=rs64+iii
here it is identified as 600Mhz RS64 IV ?

[isa-d]

My RS-64 III are 06K3016 and 06K3017, RS-64 IV are 57P5492 and 73H3708
and I'm sure of these part numbers because I personally pull out the CPUs from the boards
so seem part numbers with K are RS-64 III (part numbers changes for the speed and number of cores) and part numbers with P and H are RS-64 IV

[aberco]

1 is not a support chip, it's a dual die L3 cache (32MB?).
Check http://www.cpu-world.com/forum/viewtopic.php?t=22515&highlight=ibm+cache

I however do not know what architecture it is related to.

[Vegeta]

Thanks for the info guys!

[henriok]

Cool! I'd very much appreciate a high res photo of the RS64-III. If you could get the text on the die, that'd be awesome!

[Vegeta]

[henriok]

So I got more pictures from isa'd, and I must come to the conclusion that the middle one is a RS64-IV (I've outlined it in red in my picture).

It got smaller dies than the other RS64-III's. It's got identical capacitor layout and markings on the ceramic, as the other RS64-IV SCM, and the same size die.

[Vegeta]

They both RS64-III "IStar" and RS64-IV started in 2000 by SOI process, so they may have identical capacitor layout and markings on the ceramic? On the other hand, they are similar Part Numbers, the same as in RS64-III "06K..." May be RS64-III "IStar" and RS64-IV produced by similar 0.18 SOI process? Thus the die RS64-III "IStar" 0.18 is less than at RS64-III "Pulsar" 0.22 According to my measurements a die size 06K7145 of 132 mm2. RS64-III die size of 140 mm2, RS64-IV was 128 mm Information on the RS64-III "IStar" almost no It is known that 06K7145 is 600Mhz (John pull it from 600MHz processor board) and statistics from isa-d (all pn "06K..." is RS64-III) Although it may you are right.

Sorry for my google-translate

On photo die of RS64-IV only for info, taken from article by J. M. Borkenhagen R. J. Eickemeyer "A multithreaded PowerPC processor for commercial servers" © 2000 IBM This paper describes the microarchitecture of the RS64 IV, a multithreaded PowerPC

[aberco]

Cool, that's some very nice piece of information

The FRU on the die must help too?

I have some RS64 boards from BULL RS6000 clones that I will photograph and post here.

[henriok]

I'm sticking to my analysis that Vegeta's processor is a RS64-IV. Soley based on die size comparisons of over a dozen images of different RS64-III and RS64-IV processors.
It still might be 600 MHz, since both RS64-III and RS64-IV came in that speed.

I find it very hard to believe that two chips with different die size, transistor count, and L2 cache support should use exactly the same substrate and capacitor layout. It's much more likely that two processors of the same specifications did.