First of all I said Power5, not 975... We have no evidence of them being the same thing. What about the athlon 64 having over 200 million transistors and being 130nm... Its not that big...
Unless Power 5 has a vector unit, it will not be shipping in an Apple product. The only possible exception to this would be an Apple server and even that is so far out as to have zero possibility.
Quote:
The power5 ibm blade servers sport about 276 million transistors and are SMT. Also they are DUAL core and can run 4 threads at once. It is 130nm...
Power5 blade servers?
Any product Apple would be likely to ship would have to be functional in a desktop enclosure.
I doubt they will scrap the 970fx (if that was what you were referring to with 90nm).
Nope talking about the scrapping of the current 90nm process, which apparently they will be doing to some extent. Maybe scraping isn't the best word to choose how about heavy refinement.
Quote:
They (apple) are shipping dual 2.0gs Xserves. These are 90nm... even at 2 ghz they can be extremely useful. Can we say 2ghz g5 iMac?
I've been convinced for some time now that the 970FX was always designated for the low cost end of Apples lines. It simply does not have the cache upgrades or other improvements to take it beyond 2GHz in a high performance desktop. The imac or its replacement could certianly use this chip especially if they could hit 2.6 GHz in an iMac. If the 975 is a SMT chip this should nicely balance the performance of the line. The problem with the iMac is that it needs 2GHz as the minimal performance node.
looks great. I haven't heard anything about Hyper transport 2.0 spec yet, but having even more bandwidth sounds fantastic. now ram needs to get cheaper and hard drives faster.
looks great. I haven't heard anything about Hyper transport 2.0 spec yet, but having even more bandwidth sounds fantastic. now ram needs to get cheaper and hard drives faster.
I've never seen Architosh publish sensible (or accurate) rumours. The PowerMac uses HT as its system controller <-> I/O chip interface... accelerating it doesn't buy them much. Perhaps they are implying that the new PPC would forsake the 970's FSB in favour of an HT bus, but I don't think that's going to happen and its not clear that it would be a win. If a 9xx chip with an on-board memory controller were to show up, it might use HT as its I/O interface... in which case HT 2.0 would probably be a good idea because the PCIe controller would be moved to the I/O chip.
Comments
Originally posted by emig647
First of all I said Power5, not 975... We have no evidence of them being the same thing. What about the athlon 64 having over 200 million transistors and being 130nm... Its not that big...
Unless Power 5 has a vector unit, it will not be shipping in an Apple product. The only possible exception to this would be an Apple server and even that is so far out as to have zero possibility.
Quote:
The power5 ibm blade servers sport about 276 million transistors and are SMT. Also they are DUAL core and can run 4 threads at once. It is 130nm...
Power5 blade servers?
Any product Apple would be likely to ship would have to be functional in a desktop enclosure.
Quote:
IBM Blade servers with PPC POWER5
I doubt they will scrap the 970fx (if that was what you were referring to with 90nm).
Nope talking about the scrapping of the current 90nm process, which apparently they will be doing to some extent. Maybe scraping isn't the best word to choose how about heavy refinement.
Quote:
They (apple) are shipping dual 2.0gs Xserves. These are 90nm... even at 2 ghz they can be extremely useful. Can we say 2ghz g5 iMac?
I've been convinced for some time now that the 970FX was always designated for the low cost end of Apples lines. It simply does not have the cache upgrades or other improvements to take it beyond 2GHz in a high performance desktop. The imac or its replacement could certianly use this chip especially if they could hit 2.6 GHz in an iMac. If the 975 is a SMT chip this should nicely balance the performance of the line. The problem with the iMac is that it needs 2GHz as the minimal performance node.
Quote:
Architosh updated G5 and HT 2.0
looks great. I haven't heard anything about Hyper transport 2.0 spec yet, but having even more bandwidth sounds fantastic. now ram needs to get cheaper and hard drives faster.
Originally posted by sumoftheparts
haven't seen this mentioned anywhere yet
Architosh updated G5 and HT 2.0
looks great. I haven't heard anything about Hyper transport 2.0 spec yet, but having even more bandwidth sounds fantastic. now ram needs to get cheaper and hard drives faster.
I've never seen Architosh publish sensible (or accurate) rumours. The PowerMac uses HT as its system controller <-> I/O chip interface... accelerating it doesn't buy them much. Perhaps they are implying that the new PPC would forsake the 970's FSB in favour of an HT bus, but I don't think that's going to happen and its not clear that it would be a win. If a 9xx chip with an on-board memory controller were to show up, it might use HT as its I/O interface... in which case HT 2.0 would probably be a good idea because the PCIe controller would be moved to the I/O chip.