[naru]

Anybody have any ideas when this may appear..also when adobe might upgrade photoshop to run on the intel chip , as I need to purchase a new tower but I am loathed to buy one if a new one will be out in a month or so.

thanks

[Weezer]

first of all, a new tower desktop mac will by definition no longer by a G5 tower, it will be an intel tower. The rumors have the desktop variation (merom?) being ready to go in 6 months, so probably towers in 9 months?

[naru]

thanks for the info. I thought maybe with the new imacs and powerbook pros being out the towers wouldn't br far behind, however 6 to 9 months is too long to wait..I may have to settle for the G5 now

[Lateralus]

Agreed.

It's more likely that we'll see another revision of G5 Power Macs before the Intel towers hit.

[Todd Madson]

Especially since currently the only 64-bit offering Intel publicly has
is the Xeon processor. The new processors have to be finished first
before they can design a machine around it although Apple very likely
has pre-production samples. I'm understanding it will be later in the
year before we see an "Intel Tower" machine.

[naru]

thanks everyone..I will move forward with a dual core

[mduell]

Originally Posted by Lateralus

It's more likely that we'll see another revision of G5 Power Macs before the Intel towers hit.

I doubt it. The PowerMac's usual cycle puts them for an upgrade in about 3 months. If Apple doesn't have an Intel pro desktop Mac ready by then, I think they'll just let the current hardware run a bit longer.
I haven't seen anything from IBM about faster PPC970MP chips.

[foo2]

Originally Posted by Todd Madson

Especially since currently the only 64-bit offering Intel publicly has
is the Xeon processor.

Most of Intel's newer CPUs are now 64 bit, so I can't imagine why you'd write that. Check them out and you'll see.

[Todd Madson]

foo2: these are the chips that have been announced but not yet shipping.
I believe at this very moment, in a lab somewhere, there are production
prototypes being worked on. But its not something you will be able to buy
for a while.

[Catfish_Man]

Funny, Intel sure seems to think they're shipping 64 bit Pentium-Ds... http://www.intel.com/products/proces...um_D/index.htm

I agree, though, that Apple is waiting for a suitable chip. The Pentium-D would offer little performance advantage (if any) at fairly significant power/cost downside.

[Todd Madson]

The D doesn't compare math-wise with the 970FX PPCs or DCs unfortunately.

Pros will want more performance. They don't care so much about the
power consumption but want as much horsepower as they can get.

[foo2]

Originally Posted by Todd Madson

The D doesn't compare math-wise with the 970FX PPCs or DCs unfortunately.

Pros will want more performance. They don't care so much about the
power consumption but want as much horsepower as they can get.

This may or may not be true, but nonetheless, your original point that Intel doesn't have 64 bit chips other than the Xeons is wrong. Intel has been shipping 64 bit chips for quite some time. If Apple wanted to use them, they easily could.

[mduell]

Originally Posted by Todd Madson

The D doesn't compare math-wise with the 970FX PPCs or DCs unfortunately.

Where by "doesn't compare" you mean "is just as fast clock-for-clock, and clocks higher"? Odd way of phrasing it.

int_rate
PD: 31.4 @ 3.2Ghz (dual core)
G5: 20.2 @ 2.2Ghz (dual FX)
fp_rate
PD: 31.1 @ 3.2Ghz (dual core)
G5: 19.2 @ 2.2Ghz (dual FX)

[Lateralus]

Well Mark, the one thing I will grant you is that you're one of the biggest pro-Intel anti-PPC fanboys I've come across on the net. You even stick up for the NetBurst Pentiums. Seriously, that takes a lot of very selective reading and test gathering to be able to do.

[Salty]

Does anyone know what Intel's going to be calling the desktop chips they're coming out with? Are they also going to be branded as Core Duo and Core Solo or?

[Catfish_Man]

Originally Posted by Lateralus

Well Mark, the one thing I will grant you is that you're one of the biggest pro-Intel anti-PPC fanboys I've come across on the net. You even stick up for the NetBurst Pentiums. Seriously, that takes a lot of very selective reading and test gathering to be able to do.

Not particularly. There were large periods of time when NetBurst was the fastest available desktop processor. Even now it's still fast, it's just not fast relative to the Athlon64, or fast compared to the amount of power it draws. Looking at anandtech's benchmark's of Intel and AMD's latest and greatest ( http://anandtech.com/cpuchipsets/sho...spx?i=2668&p=7 ), the P4EE doesn't seem all that much behind the Athlon64 FX-60 until you look at the power usage chart, and I think you'll agree that the FX-60 is quite a speedy chip.

[mduell]

Originally Posted by Lateralus

Well Mark, the one thing I will grant you is that you're one of the biggest pro-Intel anti-PPC fanboys I've come across on the net. You even stick up for the NetBurst Pentiums. Seriously, that takes a lot of very selective reading and test gathering to be able to do.

I'm happy to see NetBurst go the way of the dodo, but I'll give them credit where credit is due.
It doesn't take much selective reading and test gathering to show the PD having a performance lead over the G5 for "math." Pop over to AcesHardware SPECmine and pull off the fastest result for each.

I don't really care what the instruction set is.
Building a super(cluster|computer) and need screaming fast LINPACK performance? PPC970 on the lower end and PPC440 on the higher end are clearly the way to go (top500.org). PPC970 and PPC7400 are crazy fast for RC5 (and similar) encryption routines (distributed.net). POWER5 does very well for high-end database systems (tpc.org). My EE and CE friends rave about PPC for embedded apps.
PowerPC has it's niches on the desktop; you can't build (with reasonable parts) a quad Xeon or Opteron for the price of the quad G5.
Where PowerPC falls short is laptops. Low power PPC970MP has yet to appear on any roadmap, much less on the street. MPC8641D similarly lacks presence outside the laboratories at Freescale.
I think you're seeing a bias in my comments here because most posts/threads are about laptops and slim desktops instead of superclusters, encryption performance, or embedded conrollers.

[Big Mac]

Ah the Intel fanboys strike with their SPEC benchmarks that have historically skewed in Intel's favor. Look, Yonah supposedly beats the Pentium D, and yet Yonah does not beat dual G5s, so I care little for SPEC scores.

[Catfish_Man]

Originally Posted by Big Mac

Ah the Intel fanboys strike with their SPEC benchmarks that have historically skewed in Intel's favor. Look, Yonah supposedly beats the Pentium D, and yet Yonah does not beat dual G5s, so I care little for SPEC scores.

Actually, Yonah matches the Pentium-D at SPECint_rate and loses fairly badly to it at SPECfp_rate (30ish vs. 20ish). I suggest you look a bit closer before dismissing results you don't like.

[Big Mac]

Originally Posted by Catfish_Man

Actually, Yonah matches the Pentium-D at SPECint_rate and loses fairly badly to it at SPECfp_rate (30ish vs. 20ish). I suggest you look a bit closer before dismissing results you don't like.

I told you that I don't care about SPEC benchmarks because they don't reflect real world performance. I stated Yonah seems to beat the Pentium D, but it still doesn't best the dual G5. I said nothing about Yonah versus Pentium D in SPEC. As I said, I don't care about SPEC. Take a closer look at the post to which you're responding before offering a critique.

[02gtstang]

the real question i have is will the woodcrest tower be faster than the quad g5
in fcp rendering and compressor encoding. I do a lot of HDV stuff, and as much
as I want to buy a quad, I think I may be better off waiting.

[Catfish_Man]

Originally Posted by Big Mac

I told you that I don't care about SPEC benchmarks because they don't reflect real world performance. I stated Yonah seems to beat the Pentium D, but it still doesn't best the dual G5. I said nothing about Yonah versus Pentium D in SPEC. As I said, I don't care about SPEC. Take a closer look at the post to which you're responding before offering a critique.

'k, well what I said is reflected in "real world" benchmarks as well. For floating-point heavy tasks (video encoding, for example), Yonah loses to the Pentium-D. For others it roughly equals. Very similar to how it performs vs. the G5, really (compiling, the classic branchy integer task, is absurdly faster than the G5 on Core Duo. Stuff like iDVD and iTunes encoding CD either loses or only wins by having more cores than its predecessor).

02gtstang: personally I'd get the quad. Even if Conroe lives up to the high expectations it's been generating, the quad will still be dang fast.

[t500]

According to the front page today. Adobe will not have Mac Intel apps fot another 12 months....

[Big Mac]

Originally Posted by t500

According to the front page today. Adobe will not have Mac Intel apps fot another 12 months....

And as for those who fawn over Intel, the ultimate proof will be in how long it takes Apple to kill the Power Mac line, an act which the company is apparently eager to do anyway.

[kaz7777]

according to the front page of what?

[SMacTech]

Originally Posted by t500

According to the front page today. Adobe will not have Mac Intel apps fot another 12 months....

And during this time there will be a PPC model from Apple or they will lose the Adobites to that other OS we love so much, further eroding their enterprise install base.

Now you won't see the same thing happen for the iMac which is already on the way out. I give it 3 months and it will be gone completely from the Apple store.

[mduell]

Originally Posted by t500

According to the front page today. Adobe will not have Mac Intel apps fot another 12 months....

According to Adobe's statement, they may release Universal Binaries in as little as 8 months.

[Todd Madson]

I'm still asking: how fast can that bad boy render a seti @ home block?

That's a benchmark that, though machines may be running optimized client, I'd
like to see/hear. Sadly, the guy on my team with a Pentium D machine isn't
doing too well with his:

I hate to pick on him - it's a Pentium D 3.00 ghz with 2 gigs of ram and his
stats are generally between 4000 and 8000 seconds like this:

http://setiathome.berkeley.edu/show_...hostid=1714613

I believe his hardware is a 32-bit machine though.

My own Pentium M single core machine running the Naparst optimized client
is sadly around two hours and 10 minutes to about two hours and 30 minutes.

Sadly, my G4/400 does around 2 hours 38 minutes with the Altivec optimized
client. I'd like to see the Pentium M unit do better but...

The Athlon 2400 machine (overclocked to 2.2 ghz) that I built generally does a
block in right around 2 hours even, sometimes a little less. Better.

My G5 dual 2.5 does one in about 36-38 minutes but that's running the A6 alpha
seti rendering client.

I'm told some Dual Xeons running really fast memory and the optimized client with
a really fast drive can render a block in 15 minutes. I'd like to see it because I'll
build one then. I need one for the farm.

[beldar3000]

Originally Posted by Lateralus

Agreed.

It's more likely that we'll see another revision of G5 Power Macs before the Intel towers hit.

Just remember that no one expected to see any Intel Mac until summer. I'm thinking about a new DT, but I'm fer sure waiting.