Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
New Dual Core G5 - Fall 2005
- Thread starter adamfilip
- Start date
- Sort by reaction score
its only a year away
ram hasnt changed much since the year ago when the g5 was introduced
i figured DDR2 wont be readily availale by next fall
ram hasnt changed much since the year ago when the g5 was introduced
i figured DDR2 wont be readily availale by next fall
Hate to be nitpicky but realistically, wouldn't it have a 1.75GHz FSB?
Other than that, nice specs. Let's just hope IBM can pull it off given their too recent history in production of processors. I like the external SATA connections too, could be very useful; I find the current 2 drive configuration too restricting. Oh, and STANDARD bluetooth, I got my Rev A when Rev B's were out so I couldn't customize and I'm saddened by not being able to get bluetooth without attaching a rather shoddy looking dongle to the pristine look of the G5.
Other than that, nice specs. Let's just hope IBM can pull it off given their too recent history in production of processors. I like the external SATA connections too, could be very useful; I find the current 2 drive configuration too restricting. Oh, and STANDARD bluetooth, I got my Rev A when Rev B's were out so I couldn't customize and I'm saddened by not being able to get bluetooth without attaching a rather shoddy looking dongle to the pristine look of the G5.
Blackheart said:
incase you didnt realize it.. its just made up..
anyways.. the frontside bus doesnt have to be 1/2 of the clock speed if you read the tech documents on the g5 you wil see it has other multiplier settings. aswell
similar to how the memory isnt half of the bus speed . since they are independant
but to make you happy.. now with a 1750mhz Front side bus
adamfilip said:
Yeah, I knew it was fictitious. Also, thanks for the info on the FSB speed. I just assumed that it was always 1/2 the clock speed because that's how all of the past/present G5's have been made.
Only 1MB for a dual core? What's wrong w/ them. It's a photochop! Make it unbelievable like 4MB.
I would expect faster and more dense RAM options, but otherwise it looks good to me. PC4300 DDR2 (533MHz) is already shipping in large quantity and is available as 2GB and even 4GB DIMMs. So I think it should have up to 16GB of PC4300 RAM optional (or 32GB if Apple adopts newer tech before its available in sufficient quantity). Assuming IBM fixes its problems, it looks reasonable to me.
adamfilip said:
I don't know if you'll know the answer to this, but can the multiplier basically be anything? I mean, I know the common ones like 2:1 and 3:1, but what about 7:3, which is what you'd need to get a 1.5 GHz Frontside bus on a 3.5 Ghz chip? Is that possible? What about 14:11?
You guys do realize that that fast of a Front Side Bus won't really matter unless you have ram that can match the same speed. Even the current 1.25Ghz FSB can't be fully used because steve won't upgrade to DDR2.
And yes, PC's already have DDR2 and it rawks.
And yes, PC's already have DDR2 and it rawks.
Apple will put four busses of ddr1 ram in! Then when you upgrade you have to buy 4 same sticks of ram!!
Nice idea, but what this makes me wonder is where Apple is going to go when it redesigns the current form factor.....
D
D
I would not be surprised if DDR2 made its debut in Macs along with PCI Express. Dual-Channel DDR2-600 would fill a 1.25 GHz system bus nicely.MrJohnson said:
MrJohnson said:
Wow, MrJohnson you really need to get educated on this material.
All of you need to read Memory Bandwidth vs Latency Timings. The majority of the article explains terms and then shows why DDR2 is only better in some situations; much like the P4 itself, in the pursuit of marketing leads the memory market is making questionable design choices. Bandwidth and latency are needed for different applications, and the expansion of bandwidth is leading to worse latency due to technological constraints. Simultaneously other innovations in memory bus design are making bandwidth far less important from a performance perspective by "solving" bandwidth outside the Ram stick.
All of this is fine, but page 6 is especially telling. This runaway expansion of bandwidth has been making up for its worse and worse memory timings by relying on motherboard speeds. Slower motherboard speeds in the Athlon64/XP (and by extension Macs) mean that the weak latency stats are not being made up for.
Athlons run better under tight-timing slower RAM than the currently stylish high-bandwidth DDR2 implementations, partly because the higher speed RAM can't achieve those speeds on an Athlon64; ironically the putatively slower, tighter-memory-timing sticks were able to be set to higher speeds.
Which just makes MrJohnson seem all the more foolish. Apple rightly is sticking to DDR RAM for the time being. Even speeding up the motherboard (while laudable) wouldn't change the basic validity of the DDR decision as they would no doubt further parallelize their memory bus, making the CPU latency-limited instead of bandwidth limited (which, for some applications, it already is).
Before someone asks why increasing the speed of the motherboard won't solve the latency problem as it is supposed to, on current die sizes as speed goes up the timings have to worsen, or have higher voltage throughput than JEDEC is willing to permit.
-RS
How true! DDR2's biggest issue is the increased latency it has associated with it. The way things are going right now, Mac OS X might just start using MRAM first, making DDR2 redundant...quidire said: