In facts if there is no new G4 release in MWNY : the evolution of the G4 will be finish , and the release of G5 imminent.
The release of the G5 is linked with the ability for mot to release SOI 0,13 chips. If mot can not make a DDR G4, i do not see the interest to push more the developpement of the G4, they rather switch to the G5 directly. I am almost sure that they have already developped the architecture of that chip.
Haven't seen much rumors about the G5 lately, but would be cool though *dream*
But what you're gonna get in July is a couple of speed-bumped G4's ( max 1.4 Ghz ) with new features ( seen in XServe -> 133mhz bus 133mhz DDR SDram (266 mhz), 64-bit PCI,... )
The release of the G5 will not be the death of the G4. We will see the G4 in the consumer oriented products for at least a year and probably 2-3 years after the release of the G5, whenever that is.
<strong>The release of the G5 will not be the death of the G4. We will see the G4 in the consumer oriented products for at least a year and probably 2-3 years after the release of the G5, whenever that is.</strong><hr></blockquote>
Exactly. A lot of people say "they can't release the G5 now, the G4 still has a lot of life left in it." So did the G3 when the G4 came out. It just lived in the notebooks and the consumer desktop.
Sorry that I didn't have the time to write during the last few weeks. Now I'm back and I have some further information which might be interesting.
1. Motorola shipped prototypes of the current G4 with DDR L3 cache in mid 2001.
2. Motorola produces prototypes in Q3/2001 that were very similar to the current Xserve design, using a GT64260 ASIC (that did not offer DDR support, nevertheless the Apple Xserve architecture seems to be based on this design.
3. As I mentioned in an earlier post the e500 based G5 is a 32 bit processor. As far as I can say there is no 64 bit version currently in development.
4. Some more G5 specs (MPC8540) that you probably don't know yet:
- 2-way superscalar with out-of-order issue execution and multiple execution units
- Branch unit with a 512-entry, 4-way associative branch target buffer
- One load/store unit
- Three integer units
? 2 simple units (1-cycle Ops)
? 1 complex unit for integer multiply/divide
- 1-cycle throughput with 4-cycle latency, except divide
Configurable caches
? 0-32KB I-Cache and D-Cache
? Protected by parity
? Optional on-chip L2 cache
? Protected by ECC
? External writeable
? Cache line locking in L1 caches,
L2 cache and BTB
MMU (memory management unit):
? 2 level implementation
? L1- 64-entry for 4K pages + 4-entry
for variable pages (each)
? L2- 256-entry for 4K pages + 16-
entry for variable pages (unified)
? Software Managed TLB (Level-2)
? Variable page sizes (4KB to
256MB)
? True Little-Endian Support by page
I think this makes clear that the MPC8540 will not be the G5 we will see in PowerMacs.
[quote] Well since the G5 is coming out in January, we have decided not to update the Powermacs until then. BTW we were going to show off somthing really cool at MWSF (2002), but we lost it right before the show. It realy was amazing but, well, amazing protypes don't last long at the Dry Cleaners. <hr></blockquote>
Comments
<strong>old news. never mind
[ 06-03-2002: Message edited by: soulcrusher ]</strong><hr></blockquote>
it's my bet too. But who knows ?
In facts if there is no new G4 release in MWNY : the evolution of the G4 will be finish , and the release of G5 imminent.
The release of the G5 is linked with the ability for mot to release SOI 0,13 chips. If mot can not make a DDR G4, i do not see the interest to push more the developpement of the G4, they rather switch to the G5 directly. I am almost sure that they have already developped the architecture of that chip.
Haven't seen much rumors about the G5 lately, but would be cool though
But what you're gonna get in July is a couple of speed-bumped G4's ( max 1.4 Ghz ) with new features ( seen in XServe -> 133mhz bus 133mhz DDR SDram (266 mhz), 64-bit PCI,... )
speedbumped Firewire, airport...
Speedbumped (and bigger displays) iMac, eMac
<strong>The release of the G5 will not be the death of the G4. We will see the G4 in the consumer oriented products for at least a year and probably 2-3 years after the release of the G5, whenever that is.</strong><hr></blockquote>
Exactly. A lot of people say "they can't release the G5 now, the G4 still has a lot of life left in it." So did the G3 when the G4 came out. It just lived in the notebooks and the consumer desktop.
I think the latest we'll see the G5 is MWSF 2003.
1. Motorola shipped prototypes of the current G4 with DDR L3 cache in mid 2001.
2. Motorola produces prototypes in Q3/2001 that were very similar to the current Xserve design, using a GT64260 ASIC (that did not offer DDR support, nevertheless the Apple Xserve architecture seems to be based on this design.
3. As I mentioned in an earlier post the e500 based G5 is a 32 bit processor. As far as I can say there is no 64 bit version currently in development.
4. Some more G5 specs (MPC8540) that you probably don't know yet:
- 2-way superscalar with out-of-order issue execution and multiple execution units
- Branch unit with a 512-entry, 4-way associative branch target buffer
- One load/store unit
- Three integer units
? 2 simple units (1-cycle Ops)
? 1 complex unit for integer multiply/divide
- 1-cycle throughput with 4-cycle latency, except divide
Configurable caches
? 0-32KB I-Cache and D-Cache
? Protected by parity
? Optional on-chip L2 cache
? Protected by ECC
? External writeable
? Cache line locking in L1 caches,
L2 cache and BTB
MMU (memory management unit):
? 2 level implementation
? L1- 64-entry for 4K pages + 4-entry
for variable pages (each)
? L2- 256-entry for 4K pages + 16-
entry for variable pages (unified)
? Software Managed TLB (Level-2)
? Variable page sizes (4KB to
256MB)
? True Little-Endian Support by page
I think this makes clear that the MPC8540 will not be the G5 we will see in PowerMacs.
SJ at MWNY:
[quote] Well since the G5 is coming out in January, we have decided not to update the Powermacs until then. BTW we were going to show off somthing really cool at MWSF (2002), but we lost it right before the show. It realy was amazing but, well, amazing protypes don't last long at the Dry Cleaners. <hr></blockquote>