PDA

View Full Version : G5 Altivec Carbon Fractal Benchmark




MacRumors
Jun 26, 2003, 11:15 AM
One reader submits the AltiVec Fractal Carbon (http://www.daugerresearch.com/fractaldemos/altivecfractalcarbon.html) benchmark for the 2.0GHz G5 PowerMac with SMP and Altivec enabled:

14055.2 MFLOPS

For reference, a G4/450 is 1.5 GigaFlops (~1500 Megaflops)



rice_web
Jun 26, 2003, 11:23 AM
Wait a second.... I though Altivec was toned down on the 970s.

EDIT: Nevermind, I guess it actually is.

450 x 3 (to get to roughly 1.42GHz) x 2 (for SMP) G4 = 9 GigaFlops

2000 x 2 G5 = 14 GigaFlops

So, the 970 can only crank out a linear increase despite its enormous FSB.

alia
Jun 26, 2003, 11:26 AM
What, so approximately 9 times faster then?

Alia

solvs
Jun 26, 2003, 11:29 AM
14 GigaFlops!?! Cool. Remember, the last G4 had a theoretical max of what... 18? This (if correct) is actual. And pretty good.

Edit: Wait... single or double? Wasn't the dual G4 18 (2x 9)?

Anyone know what the #s are for Intel or AMDs top offerings?

Edit again: Sorry, didn't notice the Altivec part. But still, does anyone know how many Flops the P4s and Athlons can do? (Obviously not on this, but in general)

Mr. Anderson
Jun 26, 2003, 11:30 AM
Originally posted by rice_web
So, the 970 can only crank out a linear increase despite its enormous FSB.

That seems odd, no? I looked at the link and they said that a 450 G4 (which I have) does 1.5 GFlops. That means that the Dual 2.0 G5 is a little shy of 10x the power of my machine at home. You'd think it be a little better.

Although it might depend on the application as well..

I'd love to do a Lightwave Rendering test....

D

fred
Jun 26, 2003, 11:39 AM
Well I submitted a story link to Macrumors yesterday which showed the same sorta disappointing linear type performance gains (in fact even worse) . The story is on Think Secret and the relevant quote is:


Another screenshot shows results from Skidmarks GT, a benchmarking utility that's part of Apple's CHUD performance tools. Based on the test we ran, this provides a rough comparison of G5 and G4 performance. The Skidmarks scale has "100" equal a Power Mac G4 at 1GHz. The Dual 2GHz received scores of 172 for integer performance, 270 for floating point performance, and 208 for vector performance.



http://www.thinksecret.com/news/wwdc03g5.html

hvfsl
Jun 26, 2003, 11:42 AM
Originally posted by fred
Well I submitted a story link to Macrumors yesterday which showed the same sorta disappointing linear type performance gains (in fact even worse) . The story is on Think Secret and the relevant quote is:


Another screenshot shows results from Skidmarks GT, a benchmarking utility that's part of Apple's CHUD performance tools. Based on the test we ran, this provides a rough comparison of G5 and G4 performance. The Skidmarks scale has "100" equal a Power Mac G4 at 1GHz. The Dual 2GHz received scores of 172 for integer performance, 270 for floating point performance, and 208 for vector performance.



http://www.thinksecret.com/news/wwdc03g5.html

But was the benchmark optimised for 64bit and the PPC970? Optimising ofr 64bit can make a big difference. Maybe the PPC970 does not run 32bit code as good as Apple wants us to believe. Anyway the G5 is much faster in games like Quake3, so I am ok with the G5 performance.

aasmund
Jun 26, 2003, 11:43 AM
Originally posted by solvs
Anyone know what the #s are for Intel or AMDs top offerings?


Errm did you see the name of the benchmark? Altivec Carbon Fractal Benchmark When was altivec and carbon are implemented in and on their cpu's?

nuckinfutz
Jun 26, 2003, 11:44 AM
Originally posted by fred
Well I submitted a story link to Macrumors yesterday which showed the same sorta disappointing linear type performance gains (in fact even worse) . The story is on Think Secret and the relevant quote is:


Another screenshot shows results from Skidmarks GT, a benchmarking utility that's part of Apple's CHUD performance tools. Based on the test we ran, this provides a rough comparison of G5 and G4 performance. The Skidmarks scale has "100" equal a Power Mac G4 at 1GHz. The Dual 2GHz received scores of 172 for integer performance, 270 for floating point performance, and 208 for vector performance.



http://www.thinksecret.com/news/wwdc03g5.html

How can you be disappointed about results in a Benchmark you know nothing about? Can anyone shed some light on Skidmarks? The numbers alone mean nothing. Does it accurately test Dual Processor configs?

leo
Jun 26, 2003, 11:49 AM
Was the Benchmark recompiled or even optimized for the G5?
If not, note that the Altivec code in the Fractal benchmark was handcoded with respect to the specific instruction scheduling characteristics of the G4 processor.
I wouldn't be suprised if a G5-optimized version lead to significantly better results.

jcdenton
Jun 26, 2003, 11:51 AM
Originally posted by nuckinfutz
How can you be disappointed about results in a Benchmark you know nothing about? Can anyone shed some light on Skidmarks? The numbers alone mean nothing. Does it accurately test Dual Processor configs?


I'm working on that:)

Uh, according to some mumblings <A HREF="http://www.xlr8yourmac.com/OSX/os_x_benchmark_tools.html">here</A>, Skidmarks supported only single processor tests, at least as of March 2002. The Apple Architecture and Performance Group claimed it was a processor benchmark with little system-level interaction and had a bug which occasionally posted extremely high scores.

I'm going to try and dig up some more recent informatoin, but in the meantime, I hope tihs is a good start.

BTW, if the single-processor bit is true, that means that the G5 is somewhere around twice as fast as the G4, at least in the Skidmarks test, if the numbers posted above were correct:D

Ulmer
Jun 26, 2003, 11:58 AM
Are we overlooking the fact that a dual machine isn't going to have double the benchmark score of a single machine with the same speed processor? There is going to be some loss in processing power over 2 chips right?

If a single 1 Ghz G4 gets a benchmark of 100, a dual 1 Ghz G4 machine might only get 170 not necessarily 200. Am I on the right track? This is the case in the x86 world.

So we can't really judge the dual 2.0 G5 unless we have a dual G4 benchmark to go by. Anyone have a newish dual G4, 1 Ghz or above to run the test on?

My 1 Ghz TiBook gets 3000-3500 Mflops when I run the test on it...

-Brian

jcdenton
Jun 26, 2003, 12:01 PM
Originally posted by Ulmer
Are we overlooking the fact that a dual machine isn't going to have double the benchmark score of a single machine with the same speed processor? There is going to be some loss in processing power over 2 chips right?

If a single 1 Ghz G4 gets a benchmark of 100, a dual 1 Ghz G4 machine might only get 170 not necessarily 200. Am I on the right track? This is the case in the x86 world.

So we can't really judge the dual 2.0 G5 unless we have a dual G4 benchmark to go by. Anyone have a newish dual G4, 1 Ghz or above to run the test on?

My 1 Ghz TiBook gets 3000-3500 Mflops when I run the test on it...

-Brian


Yes, you're right about the loss of processing power, although the G5 is supposed to be much more efficient in dual-processor configurations than the G4. Jobs said this at WWDC; I'm also going on information from IBM (can't remember particular links from them, but it might be mentioned in the PowerPC 970 reports over at www.arstechnica.com).

fred
Jun 26, 2003, 12:03 PM
Originally posted by Ulmer
Are we overlooking the fact that a dual machine isn't going to have double the benchmark score of a single machine with the same speed processor? There is going to be some loss in processing power over 2 chips right?

If a single 1 Ghz G4 gets a benchmark of 100, a dual 1 Ghz G4 machine might only get 170 not necessarily 200. Am I on the right track? This is the case in the x86 world.

So we can't really judge the dual 2.0 G5 unless we have a dual G4 benchmark to go by. Anyone have a newish dual G4, 1 Ghz or above to run the test on?

My 1 Ghz TiBook gets 3000-3500 Mflops when I run the test on it...

-Brian

Yes you are correct in your assumptions.....but what I find disconcerting is that according to the pre-970 hype these chips were supposed to be orders of magnitude superior to the G4 and that now seems to be a pipe dream...

Dont Hurt Me
Jun 26, 2003, 12:28 PM
we wont know for sure until the real machines are in the hands of the people and then we can run some benches, i guess this means sept2 until we can do this. i wish apple would have shown us how UT2003 runs on a 1.6, 1.8 and duallies.

jcdenton
Jun 26, 2003, 12:46 PM
Originally posted by nuckinfutz
How can you be disappointed about results in a Benchmark you know nothing about? Can anyone shed some light on Skidmarks? The numbers alone mean nothing. Does it accurately test Dual Processor configs?


I'm working on that:)

Uh, according to some mumblings <A HREF="http://www.xlr8yourmac.com/OSX/os_x_benchmark_tools.html">here</A>, Skidmarks supported only single processor tests, at least as of March 2002. The Apple Architecture and Performance Group claimed it was a processor benchmark with little system-level interaction and had a bug which occasionally posted extremely high scores.

I'm going to try and dig up some more recent informatoin, but in the meantime, I hope tihs is a good start.

BTW, if the single-processor bit is true, that means that the G5 is somewhere around twice as fast as the G4, at least in the Skidmarks test, if the numbers posted above were correct:D

Bob Dobbs
Jun 26, 2003, 01:03 PM
i'm getting about 5500 on a dual 1 gig with 512 ram - also have lots of apps open - dont know if that matters

Dave K
Jun 26, 2003, 01:10 PM
Originally posted by jcdenton
Yes, you're right about the loss of processing power, although the G5 is supposed to be much more efficient in dual-processor configurations than the G4.

The G5 benefits from not having to share a single FSB between two chips. Otherwise, it's still subject to the rules of SMP scaling which usually state roughly 80% as the typical performance increase when adding a second proc.

Which makes a dual 2Ghz G5 roughly equivilent to a 3.6 Ghz G5 when things are SMP aware. (How the dual having 2Ghz of bandwidth between two procs compared to the 1.8 Ghz a 3.6 would have affects overal performance between the two setups, I'd have to leave to someone else to sort out.)

Bob Dobbs
Jun 26, 2003, 01:11 PM
hmmm not quite as smoking as i hoped

fred
Jun 26, 2003, 01:37 PM
On the other hand... I just read this on Accelerate Your Mac which augurs well:

"Hey everyone, I've been at Apple's developer conference and had a chance to install and try out After Effects on a new G5.
I ran the Night Flight file that has come to be the standard for AE benchmarking. Since I didn't want to sit there and watch it render for hours, I ran just the first 10 interlaced frames from the project's pre-set render queue...
http://www.aefreemart.com/tutorials/3DinAE/nightflight/nightflight.html
Here are my results for this test on the three computers I have available to me:

1 x 1.0 GHz G4 PowerBook 17" - ~30 minutes (3 min/frame)
2 x 2.66 GHz Pentium Xeon from Boxx - 11 min, 39 sec (1.2 min/frame)
2 x 2.0 GHz PowerMac G5 - 6 min, 1 sec (0.6 min/frame)

I ran the Xeon test on a couple different identical machines to make sure mine wasn't just running slowly, but got identical results.
Of course my Mac bias is well-documented, but I'm sure many people here can vouch for me as an honest person. If the results had gone the other way, I'd just keep my mouth shut and let someone else break the bad news.
Other observations about this test that may ultimately work in the Mac's favor:

1) The machine was not running 64-bit Panther, but only a tweaked version of 32-bit Jaguar. Likewise, AE is obviously not yet compiled to take advantage of the G5 chip in any way. Both or these situations will automatically be rectified in the future.

2) Night Flight is very CPU-intensive, but not very disk I/O intensive. I think the 1 GHz system bus and other details on the G5 will provide greater gains for typical projects that rely more heavily on I/O."

jayscheuerle
Jun 26, 2003, 01:48 PM
Originally posted by hvfsl
Maybe the PPC970 does not run 32bit code as good as Apple wants us to believe.

Of course not. Nothing about the G5 will be as good as Apple wants us to believe.

Steve Jobs is the king of hyperbole. He tells a great fish-story and is wonderful at getting the troops amped, but you'd have to have a serious long-term memory problem to take his statements at face value.

That's not to say it isn't a welcome, but long-coming upgrade. Of course it is, but it's not going to cook you waffles, mow your lawn and save your marriage like the RDF marketers would have you believe.

neutrino23
Jun 26, 2003, 01:50 PM
Originally posted by fred
.....but what I find disconcerting is that according to the pre-970 hype these chips were supposed to be orders of magnitude superior to the G4...

As I recall, from the bench marks released by IBM the 970 was expected to be 2 to 3 times faster than a G4 for some tasks (FP?), not orders of magnitudes.

Also, performance generally doesn't scale linearly with clock speed. Look at these Xbench scores.

Dual 1GHz G4 about 100
Dual 533MHz G4 about 75

We may need a new version of xbench to get accurate scores for the G5.

jcdenton
Jun 26, 2003, 01:52 PM
Originally posted by Dave K
The G5 benefits from not having to share a single FSB between two chips... Which makes a dual 2Ghz G5 roughly equivilent to a 3.6 Ghz G5 when things are SMP aware. (How the dual having 2Ghz of bandwidth between two procs compared to the 1.8 Ghz a 3.6 would have affects overal performance between the two setups, I'd have to leave to someone else to sort out.)

Thanks for the clarification, Dave.

Frobozz
Jun 26, 2003, 02:06 PM
Originally posted by fred
Well I submitted a story link to Macrumors yesterday which showed the same sorta disappointing linear type performance gains (in fact even worse) . The story is on Think Secret and the relevant quote is:


Another screenshot shows results from Skidmarks GT, a benchmarking utility that's part of Apple's CHUD performance tools. Based on the test we ran, this provides a rough comparison of G5 and G4 performance. The Skidmarks scale has "100" equal a Power Mac G4 at 1GHz. The Dual 2GHz received scores of 172 for integer performance, 270 for floating point performance, and 208 for vector performance.



http://www.thinksecret.com/news/wwdc03g5.html


Would I be correct, or incorrect, to say that these tests do not utilize the hard drives or memory bandwidth. In that sense, they are pretty poor comparisons. The bottom line is that the architecture of these new machines should prvide a more promosing future for rapid GHz increase, and an overall speed improvement.

I'm not shocked by the scores at all.

BeigeUser
Jun 26, 2003, 02:21 PM
Back to original topic, the G5 has been scaling more than linearly.

(Assuming dual processors to be 1.8 times the performance of single processors.)

2000MHz/450MHz= 4.444

1500MFlops x 4.444=6666.666 Theoretical 2GHz rate if the CPU scaled linearly

6666.666 x 1.8 = 12000 Theoretical Dual 2GHz linear rate

If the Dual 2GHz G5 scored 14055.2, that's not so shabby.

And that's comparing to the G4/450 which didn't have a system bus bottleneck. Later G4 models had some serious bandwidth problems. Someone wrote that their Dual Gig G4 scored 5500. The Dual 2GHz G5 almost tripled that number despite only being double the clock rate.

BTW, I remember reading somewhere that After Effects is not dual processor aware and not altivec enhanced.

PaisanoMan
Jun 26, 2003, 02:39 PM
Originally posted by neutrino23
As I recall, from the bench marks released by IBM the 970 was expected to be 2 to 3 times faster than a G4 for some tasks (FP?), not orders of magnitudes...

Yeah, and even that's pretty impressive on paper (clock-for-clock). I think the expected "order of magnitude" came from the snowballing hype in these forums surrounding the new chips.

I think the most significant upgrades in the new PowerMac are ... everything. Apple's introduced a completely new foundation for their PowerMac line (not sure how the PCI-only variant fares), which should yield improved performance in every category -- disk (slightly faster with SATA), memory (both capacity and speed), CPU, I/O (depending on their super-south bridge).

I'm not surprised that the G5 "scales linearly" in CPU-level tasks ... but I think it's okay, since we probably would never have seen a 2.0 GHz G4 to test it against. ;)

fred
Jun 26, 2003, 02:50 PM
Originally posted by PaisanoMan
Yeah, and even that's pretty impressive on paper (clock-for-clock).

Impressive compared to what ??? Given the fact Apple's so far behind in the clock speed war, yeah it's impressive...overall a linear scaling is quite disappointing given they're only at 2 Gigaherz and charge a hefty premium for their machines.... Let's hope that real world performance makes up for things...and by all indications that seems to be the case as the overall architecture seems to kick serious butt what with the impressive bus and memory bandwidth these machines display ...


I'm not surprised that the G5 "scales linearly" in CPU-level tasks ... but I think it's okay, since we probably would never have seen a 2.0 GHz G4 to test it against. ;)

Not so sure....Motorola has a 2 Gig chip scheduled to come out which consumes very little power

ffakr
Jun 26, 2003, 03:06 PM
one point that I haven't seen mentioned yet. At the keynote, Adobe said that photoshop will be optimised for the 970 later this year.

The performance shown so far has been VERY impressive. The machines are running 10.2.7 which is essentially a hack to get them running for the show.

Apple is still optimising their dev tools for the G5, they don't have an OS that is optimised for it yet.. and the apps they've been demoing aren't optimised for the new architecture. The performance stats will only go up from this point.

Personally I haven't had a chance to benchmark the G5 yet, but I have played with one. They have a pretty bland out of the box software install on them so I didn't get to push it too much. My observations:
* it has a long pause before booting (screen black)... possibly 5 seconds.
* it spins the curson on the gray screen for a long time as if fsck'in the disk. Not sure why. This was about 10 seconds.
* when the blue OS boot screen comes up, it loads the system and all deamons in about TWO TO THREE seconds. It's startlingly fast.
* finder loads in under two seconds.
* traditionally slow loading apps like iPhoto and iMovie load in one Dock bounce.

I was unable to find any way to slow the machine down. Unfortunatly they didn't have anything interesting on the box I played on... no sample iMovie projects, no Mathematica...
The machine I was on was not hard wired to the network and that room was not getting wireless reception (another thing I noticed, changes to the Airport were instantaneous, much unlike my ibook which pauses)

The boxes are obviously pre-production demo models. Some had issues where the fans were not behaving as they should (spinning up for no reason...)
My machine went to sleep before I sat down at it and wouldn't wake up.. It needed to be power cycled.

Again, everyone needs to remember that these demos and benchmarks are done with tools compiled with compilers that don't know anything about the G5. Other tests are running on a Kludge of an OS... and it's all being done on demo hardware.

You'll have to trust me. The G5 is VERY fast, but we won't know just how fast for another month.
You won't be disappointed.

illumin8
Jun 26, 2003, 04:39 PM
Originally posted by ffakr
Some had issues where the fans were not behaving as they should (spinning up for no reason...)
I think this is due to the way they are controlling the fans to keep noise levels down. I work on Sun servers, and on our higher end servers, this is exactly what we do.

Basically, it works like this:

1. When the box is first powered on, the fans come on at maximum speed to ensure that nothing will overheat.

2. Once the OS is loaded, it starts an "environment monitoring daemon" that monitors temperatures of key components and adjusts the fan speed appropriately. (this is very scary the first time it happens, because it sounds like the fans just shut off, but actually they are just slowing down to a better level)

3. If the environment daemon crashes or stops running for some reason, the fans all spin back up to maximum speed just to make sure that nothing overheats. This is a failsafe method to make sure nothing will burn up.

I'm 99% sure this is how Apple is managing the heat on these boxes while making sure that everything stays quiet.

Another thing to consider with this is that variable speed fans tend to wear out a lot quicker than fixed speed fans because the speed is always changing. I would HIGHLY RECOMMEND purchasing Applecare if you buy one of these G5s because with 9 fans I'm sure you'll have at least 1 or 2 fan failures by the time you've had the box for 3 years or so. Hopefully the environment monitoring daemon will notify you of a fan failure and shut down the box if necessary. Most likely, the other fans will spin up to full speed if a failure occurs and keep the box running cool.

newfeet
Jun 26, 2003, 05:15 PM
I agree that many of these benchmarks are not optimized for the G5 (or for 64 bit) and are also being run on a 32-bit operating system 10.2. Until both the benchmarks and the operating system is optimized, the results can't be trusted completely.

What does anyone care what the benchmarks say, anyway, when the real-world tests (Mathematica, Photoshop, etc) are what matter. Who makes money rendering fractals? Personally, I'm not disapointed in the least at these speed results.

luiss
Jun 26, 2003, 06:43 PM
Remember that the G5 system's greatest advantage is it's bandwidth. These tests are pure CPU intensive.

SuzanneA
Jun 26, 2003, 08:57 PM
Originally posted by illumin8
I think this is due to the way they are controlling the fans to keep noise levels down. I work on Sun servers, and on our higher end servers, this is exactly what we do.

Basically, it works like this:

1. When the box is first powered on, the fans come on at maximum speed to ensure that nothing will overheat.

2. Once the OS is loaded, it starts an "environment monitoring daemon" that monitors temperatures of key components and adjusts the fan speed appropriately. (this is very scary the first time it happens, because it sounds like the fans just shut off, but actually they are just slowing down to a better level)

...


Oddly enough, this is exactly how my MDD dual G4 seems to work. When I power on the fans are running full speed, and LOUD. About 1/4 of the way into the boot, they sound as if they just stop, when in fact they are just slowed down. If I do something REALLY intensive (which is surprisingly hard to do, but divx encoding can do it) they start to become a bit more noisy again.

I have never, however, gotten them back to the inital boot level.