Talk:IBM z196

Great

Great Now IBM has reentered CPU Business. Even Intel and AMD cannot Challenge It. Wish to see IBM CPU in Normal, Househould Computers also. This will drasticaly bring down CPU price and also Kick innovations. —Preceding unsigned comment added by 120.61.15.133 (talk)

IBM never left. They are designing CPUs like crazy but not x86 processor that are probably familiar to you. They produces a variety of Power Architecture lines of processors for embedded systems (set top boxes, networking, storage), high-end servers, super computers and game consoles (Wii, Xbox 360 and PlayStation 3), as well as highly customized and proprietary mainframes (like this one).

This will not bring down prices as the cheapest computer with this particular computer _begins_ in the $100.000 range. –– Henriok (talk) 13:31, 27 August 2010 (UTC)

hey buddy Henriok, I guess you have not done your history class properly. remembet power pc was led by IBM for personal computing but failed due to introduction of Intel based x 86 processor and couldnot match to intels products. —Preceding unsigned comment added by 120.61.2.165 (talk) 07:42, 24 September 2010 (UTC)

I can assure you that there's nothing wrong with my knowledge in CPU history in general or PowerPC history specifically. There's a LOT of CPUs that doesn't get used in personal computers. Actually, most don't, and in those segments x86 processors are rare, easily dominated in numbers by ARM, PowerPC and MIPS processors. In this particular segment (mainframe computers), IBM has been cranking of processors massively fast, complex, capable and insanely expensive since the last 40 years or so. This is just the latest iteration of that process. Nothing has really changed, so this won't have any bearing on the CPU business at large. And again: IBM has never left it. It's just that most people only know and think of x86 processors from AMD and Intel, and IBM left THAT segment a while ago. This processor certainly won't change that. -- Henriok (talk) 09:45, 24 September 2010 (UTC)

z196 not a microprocessor

The term z196 refers to the entire box, not to an individual component. Shmuel (Seymour J.) Metz Username:Chatul (talk) 16:10, 30 August 2010 (UTC)

Well, the "z196 (microprocessor)" refers to the CPU, the individual component while "zEnterprise 196" refers to the entire box. I sympathize with your statement though and would readily change the name of this article if there was a better name for it, but IBM doesn't seem to be so forthcoming. I will clarify the article. -- Henriok (talk) 22:14, 30 August 2010 (UTC)

Much of the material has nothing to do with the microprocessor, but rather to the entire system. It really belongs in IBM zEnterprise System, and the part that would be left wouldn't be large enough to justify a separate article. I propose merging. Shmuel (Seymour J.) Metz Username:Chatul (talk) 14:45, 31 August 2010 (UTC)

• Disagree with merger. A properly researched and sourced article on the zEnterprise 196 central processor would be able to stand on its own. The same would be said for the IBM zEnterprise System article. I think both articles could be comparable to the pairing of IBM System z10 and IBM z10 (microprocessor). It's a matter of someone taking the time to work on the articles properly. I'm willing to give it an attempt before the just started semester gets too busy with my research and other activities. Also, a note to the merge nominator, you didn't provide the correct article name on the face of this article and you failed to complete the other half of the merge nomination (a merge template on the page where you are targeting the merge to go to). --TreyGeek (talk) 15:16, 31 August 2010 (UTC)
• Disagree with merger. Completely different focus and domains of knowledge. Lots of examples of similar division. -- Henriok (talk) 17:55, 31 August 2010 (UTC)
• A stand-alone article called, e.g., IBM zEnterprise System design, IBM zEnterprise System implementation, IBM zEnterprise System technology, with this content would be reasonable. However, much of the material has nothing to do with the title, which was why I suggested a merger. Further, there are statements that are true for the zEnterprise System but false for the microprocessor, e.g., the claimed number of new instructions.

  A move and some rewording would satisfy my issues as well as a merger, if that is acceptable. Shmuel (Seymour J.) Metz Username:Chatul (talk) 13:01, 3 September 2010 (UTC)

Now that I finished a rewrite of the IBM zEnterprise System article I plan on working on a rewrite of this article. My plan is to have the article focus on the details of the chip itself (not the overall system and not deeply into the instruction set which are different articles.) It'll take me a few days to do all of the research so that everything is properly sourced. So we'll see if your comments are addressed when I get that done. --TreyGeek (talk) 14:12, 3 September 2010 (UTC)

Much of this and what's been discussed on the 10 article could/should really go into the z/Architecture article. No? -- Henriok (talk) 18:17, 3 September 2010 (UTC)

The implementation details don't belong in an architecture article, and there's enough material that should be added that a merger would make the article too large. It might be reasonable to have implementation articles covering multiple models. Shmuel (Seymour J.) Metz Username:Chatul (talk) 20:08, 3 September 2010 (UTC)

Do you have any source of information for the correct nomenclature of the processor chip? Shmuel (Seymour J.) Metz Username:Chatul (talk) 20:08, 3 September 2010 (UTC)

Unless there is a sudden outpouring of support for the merge, I'll be removing the merge templates about this time tomorrow. --TreyGeek (talk) 19:57, 8 September 2010 (UTC)

L4-Cache an more

In the article there are two numbers: 196 MB (Description) and two times 96 MB (Storage Control).
What is wrong? I think in the Description it should be 192 MB.
And last: how can 1 z196 can have a total of 376 MB cache (L1-L4)? 4 x (64 KB + 128 KB + 1.5 MB) + 24 MB + 192 MB = 6.75 MB + 24 MB + 192 MB = 222.75 MB
Only for a MCM 376 MB Cache is correct: 6 z196 chips contain 6 x 30.75 MB = 184.5 MB. Plus L4 cache of 192 MB = 376.5 MB
I changed the 196 MB to 192 MB in Description. --Kiu77 (talk) 11:42, 5 September 2010 (UTC)

Second thoughts: Important is the amount of cache to which 1 core can have access. In 1 MCM this is L1 = 64 KB + 128 KB; L2 = 1.5 MB ; L3: 24 MB; L4 = 192 MB. I don't believe that a core can have access to the L1-L3 cache of other CPUs. Therefore total usable cache of 1 core is: internal 25.75 MB + external 192 MB = 217.75 MB. Not accessable caches don't count in my eyes.--Kiu77 (talk) 11:55, 5 September 2010 (UTC)