Match these G5 Specs

imgod2u · Sep 16, 2003

This benchmark from NASA comparing an un-otpimized SINGLE G5 says otherwise.

http://members.cox.net/craig.hunter/g5/

I hope you realize that 2x the processor != 2x the speed. In fact, in a single-threaded application such as the one tested in your post, it will not make a difference at all. Hence, the single 3.06 GHz P4 (your tests showed the 2 GHz G5 similar to a 2.66 P4) would be faster than a dual 2 GHz PPC 970 in that benchmark.
Other programs differ and may scale better when going to dual processors, but then again, other programs may run better on the P4 or vice versa.

Take Lightwave,
<A HREF="http://www.blanos.com/benchmark/" target="_new">This site</A> has several submissions for different Lightwave scene rendering times. Unfortunately, there was only 1 submission for a dual PPC 970 2 GHz (G5) system for the Tracer Radiosity benchmark. It took 550 seconds to render vs 395 seconds for a 3.06 P4 submitted.

It's pretty difficult to find more relavent (Photoshop filtering, video encoding, gaming, not that there are many games for Mac anyway) benchmarks as few have been allowed to benchmark these new dual G5 systems (common Apple, can't spare one machine to a review site?) and posted them on a website.

However, judging from the one you've posted and the lightwave rendering mark, the PPC dual 2.0 G5 can probably perform similarly to a single 3.06 P4 (in any decently multithreaded benchmark). In single-threaded applications, it'd be no contest, the P4 would be ahead.

"We are Microsoft, resistance is futile." - Bill Gates, 2015.

zejs1 · Sep 16, 2003

"Specs and prices, please, that tells me nothing about what the Dell has....facts, please, facts."

www.dell.com

Crashman · Sep 16, 2003

I AM a factual source.

I forget I'm talking to an idiot/newbie/mackie/whatever (doesn't know crap about PC pricing). I'm a system builder. That means I'm a source. That meanse I'm a source for prices also. Those number are generous enough to ensure that I make at least SOME money in the build. Go to Newegg and put together a price list, you'll understand then.

Most of those parts are cheap, only the CPU is expensive with the P4's, and only the motherboard is expensive with the dual AMD.

Only a place as big as the internet could be home to a hero as big as Crashman!
Only a place as big as the internet could be home to an ego as large as Crashman's!

shadow_of_colors · Sep 16, 2003

Model Specifications

Dual AMD® Opteron® Processor Capable
AMD-8131 HyperTransport PCI-X Tunnel
AMD-8151 HyperTransport AGP Tunnel
128-bit Dual Channel Memory Bus
Up to 16GB ECC Registered 333MHz DDR
(8) DIMM Slots
Dual Channel UltraDMA 133 IDE Controller
Four Channel Serial ATA RAID (0,1,0+1)
6 Channel Audio
(1) 8x AGP Pro Full Length Slot
(2) Full Length PCI-X 64bit/133MHz Slots
(2) Full Length PCI-X 64bit/66MHz Slots
(1) Onboard 10/100/1000Mbs Ethernet Adapter
(4) USB Ports: (2) Front USB 1.1, (2) Rear USB 2.0
(2) IEEE1394 Ports: (1) Front, (1) Rear
(4) 3.5" x 1" Internal hard drive bays
(3) 3.5" Exposed drive bays
(2) 5.25" Exposed drive bays
460W Power supply
(2) 92MM, (1) 80MM Cooling fans
1.44MB Floppy drive
Tower Chassis (Optional 4U Rackmount kit)
Physical dimensions: 7.0"W x 17.0"H x 17.5"D

Quote Number: 377987
Monday, September 15, 2003

Configuration

1 x 200GB 7200 rpm ATA133 hard drive 8MB Cache
1 x DVD+R/+RW/-R/-RW CDRW Combo drives EIDE (black bezel)
1 x 3 Button Kensington PS2 Mouse
1 x Dual Opteron Model 240
1 x 1GB PC2700 ECC Reg DDR (2- 512MB DIMMS)
1 x 1GB PC2700 ECC Reg DDR (2- 512MB DIMMS)
1 x NVIDIA Quadro4 380 XGL 64MB 8X AGP
1 x Windows XP Professional Edition
1 x Black 104 key keyboard
System Total: $3,441.00
Accessory Total: $0.00

*Grand Total: $3,441.00

from BOXX technologies. that's a Dual AMD Opteron work station. www.boxxtech.com if you want to go build your own. This is extreme overkill for a PC, since it is a workstation. the Stats on the G5 are more of a workstation quality system, not a home PC. plus, the G5 doesn't even operate in 64-bit mode last i checked. (i could be wrong, i have been before, and will be again, so i'm not afraid to admit it) the Opteron won't either unless you're running the specific OS for it.
anyway, my two cents into the arena
Cheers

Crashman · Sep 16, 2003

Hehe, just like I said, only this one's better, the link you gave shows that 1 2.66GHz P4 is barely faster than a 2x2GHz G5.

And a 3GHz P4 is even faster still. Thanks, now I can save even more money in the comparison and rate it by 2.66GHz P4's. But that seems kind of pointless since the 2.60GHz is faster and actually makes use of DDR400 in dual channel mode.

Now go to Newegg and price it out, the cheapest DDR400 (PC3200), the cheapest Radeon 9600 Pro, a retail boxed P4 2.60GHz, etc.

Only a place as big as the internet could be home to a hero as big as Crashman!
Only a place as big as the internet could be home to an ego as large as Crashman's!

tombance · Sep 16, 2003

You do realise that the P4 2.66Ghz surpassed any Mac processor in two of those tests. The other one in whuch the P4 was involved was basically IPC, where it was beaten by all of the Macs, but IPC only means anything if the processors are going to be running at the same speed, and theyre not.

<A HREF="http://service.futuremark.com/compare?2k1=6752830" target="_new">Yay, I Finally broke the 12k barrier!!</A>

pitsi · Sep 16, 2003

I have this small test for you. Let's say you are a home user and need to buy a new system tomorrow. Now, would you prefer the G5 with those specs you mentioned for $2999 or the following system?:

• Pentium 3.0c
• Asus P4P800 Deluxe
• 512MB PC3200 DDR SDRAM
• 128MB Radeon 9800 Pro
• Sony DVD+-RW Drive (DRU-510A)
• Seagate 160GB SATA HDD
• On-board 10/100Mbps Ethernet
• 2 x Firewire Ports
• 6 x USB 2.0 Ports
• Creative Labs Sound Blaster Audigy 2
• Antec Sonata Mid-Tower Case

Total: $1500 (!!!)

Half the price! And my point is this. Yeah, maybe the G5 is suitable for SOME people that work with specific apps where the G5 is good at. But for the rest of us (and I mean 98%), I believe the system I just posted would suit us better .. for $1500 less !!!

LancerEvolution7 · Sep 16, 2003

I constantly hear how the G5 is overpriced, so I thought I'd put up a challenge. Can anyone match these specs with a matching dual 64bit PC system? Although the Build Your Own crowd is welcome to put their two cents in, I think it would be more fair to configure an factory-supplied configuration.

So what you are telling me is paper numbers are better than actual speed? Tell me, what do you use on you G5 Mac. I can give you a PC that is 1/2 the price and will perform better than that mac.

No, for a couple of reasons.

32-bit systems can only support 4gigs of RAM, and once Panther (OS X.3) ships and programs are optimized for 64bits, 32bit systems will not out-perform 64bit systems (not to mention the fact that it is highly debatable whether any 32-bit system right now could outperform this configuration, see the following link for a fair and objective look at SINGLE G5 benchmarks...as it says in the study "Note that even though the G4 and G5 systems have dual processors, detailed benchmarks in the present study pertain to a single processor only.")

Again more paper. First of all you probably won't need 4gigs of Ram until maybe one year later if that soon. Most people can't even use 1 gig now. Second, 64 bit programs will be comming out SOON. So even if it does outperform 32bit processors, we don't know by how much. Third, there aren't many 64bit programs right now. So we can worry about 64bit programs when it gets widespread. A64 will be comming out soon also.

<A HREF="http://www.barefeats.com/pentium4.html" target="_new">Benches of P4 VS G5.</A>

The only debatable thing apple has better than PC is the OS which is SUPPOSED to be more stable. Although WinXp and 2K is extremely stable, but that goes into another debate.

slvr_phoenix · Sep 16, 2003

This benchmark from NASA comparing an un-otpimized SINGLE G5 says otherwise.

http://members.cox.net/craig.hunter/g5/

Yes, that link shows exactly how well the G5 stacks up to those old 32-bit PoS from Intel. I mean the dual 2GHz G5 only lost to the single 2.66GHz P4 by 1 MFlop. Why that's a veritable tie if I ever saw one! Great job IBM and Apple!

Oh, wait ... that was a single P4 (NorthwoodB) 2.66GHz running single-channel PC2700. Gee, I wonder how well a P4 Northwood C with it's improved FSB bandwidth and and HyperThreading would do. A P4 NWC 2.4GHz generally ties with a P4 NWB 2.66GHz in performance.

And then what if we actually used a modern motherboard and dual-channel PC3200 instead of single-channel PC2700? I mean the Pentium4 is a very bandwidth-hungry processor and performs rather poorly when it's bandwidth needs aren't being met. So that should also really boost performance.

And crap, even the Northwood C is up to 3.2GHz, which is a considerable improvement over that 2.66GHz as well.

But nah, there's no way that the P4 (<A HREF="http://www.supermicro.com/PRODUCT/MotherBoards/E7505/X5DAL-G.htm" target="_new">or dual 3.06GHz Xeons with 8GB of RAM</A>) could possibly outperform that G5. Apple still has it in the bag. Yeah. Really. Honest.

<pre><A HREF="http://ars.userfriendly.org/cartoons/?id=20030905" target="_new">People don't understand how hard being a dark god can be. - Hastur</A></pre>
* edit: Made the 8 in 8GB really stick out.Edited by slvr_phoenix on 09/16/03 12:49 PM.

pjordan · Sep 16, 2003

Can the G5 be overclocked? The 2.4C can be overclocked to 3.0. It can have 1100 mhz Quad FSB. Costs 180 for CPU, 100 for motherboard, and like 210 for memory. Total $390! Course you would want a few other things. Would beat the G5 most every way.
Imagine getting 10 of those and hooking them together...
The price would be comparable.

pIII_Man · Sep 16, 2003

MFlops should not be effected much if at all by increased memory bandwith....

Proud owner of DOS 3.3

slvr_phoenix · Sep 16, 2003

MFlops should not be effected much if at all by increased memory bandwith....

Tell that to the scientists who crunch through large streamed data sets, the folks who do MPEG compression/decompression, and the artists who crunch through intense 3D renderings.

It all depends on the application. The program that was benchmarked may or may not be affected much, but there is definately software that uses floating-point precision that is.

Besides, that was only one point of many. 😉

<pre><A HREF="http://ars.userfriendly.org/cartoons/?id=20030905" target="_new">People don't understand how hard being a dark god can be. - Hastur</A></pre>

pIII_Man · Sep 16, 2003

I know this, i ment most MFlop benchies would not detect the diffrence and MFlops mean in the real world, unless you are doing a purely cpu intensice app.

Compare a P4 NW A to NW C in performance

Proud owner of DOS 3.3

Edited by piii_Man on 09/16/03 04:07 PM.

slvr_phoenix · Sep 16, 2003

I know this, i ment most MFlop benchies would not detect the diffrence and MFlops mean in the real world, unless you are doing a purely cpu intensice app.

I'm not sure that I follow what you're trying to say.

Compare a P4 NW A to NW C in performance

Okay. There's a huge performance difference at the same clock across the board. Even if you look at fp-specific benches like SPECfp you get:
2.8GHz 4x100MHz FSB = 878 / 887 = 74% of C
2.8GHz 4x133MHz FSB = 1010 / 1020 = 85% of C
2.8GHz 4x200MHz FSB = 1186 / 1204 = 100% of C
So from A to B you get an 11% jump. From B to C you get a 15% jump.

So I'm just really confused. The FSB makes a huge difference in a P4s performance, whether in Flops or Iops or whatever.

<pre><A HREF="http://ars.userfriendly.org/cartoons/?id=20030905" target="_new">People don't understand how hard being a dark god can be. - Hastur</A></pre>

pIII_Man · Sep 16, 2003

I always beleived that MFlops were purley cpu horsepower related, as in, not effected noticably by things like bus speed, cache and memory timmings. Hehe, looks like i was wrong...

Proud owner of DOS 3.3

addiarmadar · Sep 17, 2003

That G5 is extremely impressive and can smoke any PC out in the market but wat makes it or in generally expensive is tha lack of software you can get for them. PCs always get the first dibbs on software.

F-DISK-Format-Reinstal DO DA!! DO DA!!

LancerEvolution7 · Sep 17, 2003

Umm, it'd be nice to back it up with some proof.

imgod2u · Sep 17, 2003

I always beleived that MFlops were purley cpu horsepower related, as in, not effected noticably by things like bus speed, cache and memory timmings. Hehe, looks like i was wrong...

The question always is, how are you going to measure it? You need data to work those instructions on and you have to get that data from somewhere. Guess where? Main memory. There's no way to measure "pure CPU" because the processor *always* depends on memory. Now, you can try your best to minimize the impact memory has, such as write your code to fit in cache most of the time, preload data.
All in all, at the end, memory will still play a role no matter how hard you try to rule it out.

"We are Microsoft, resistance is futile." - Bill Gates, 2015.

cdpage · Sep 17, 2003

I remember back a couple years when my freind picked up a dual G4 800 or 866(not too sure on that spec. It is a dual 8 something any how.

We both did alot of work on stupid iMacs at the school... these were only G3 400's and 500's So as you may have figured the differace beyween them were Massive.

however we did do some testing (simple stop watch method)
with a couple filters.

our fastest iMac 500 we could find ( as school computers... PC or MAC are inherently CRAP )
I on my P1 233
him on his G4 800 MP
another friend on a P3...or P2, (sorry i forget) 733

as you can see it was a while ago, and we obvioulsy didn't wright anythng done case that would have been a good idea.

andhow, his G4 creamed all our machines...what ever the numbers were i remember that the iMacs were about 2x faster then mine the 733 was about the same or little faster then the iMac and the G4 was well over 2x faster then the iMac.

I do remeber a few things here and there about how many poeple talked about and asked question of what was realy faster a LOT. Most of the time people came up with per MHZ the Mac is faster.

That little gain i suppose is the price that one had to pay for the "better" machine. but since the PC (intel & AMD) are now more then 1Ghz faster (150%) in the Hz department, that little bit faster per/hz thing that they have, can't compete.

so i have NO proof here, but a handfull of students who are all now Graphic designers, some on PC some on MACs. can attest to the per/hz thing.

I'd say its a good split 50 50 of what they have at home AND for those that go to work use its a little more on the PC side in the GD field... 40 60.

ASUS P4S8X - P4 2.4B - 2 x 512M DDR333 - ATI 9500 Pro(Sapphire) - WD 80G HD (8M Buffer) - SAMSUNG SV0844D 8G HD - LG 16X DVD - Yamaha F1 CDRW - Iomega Zip 250 int.

cdpage · Sep 17, 2003

Ok talked to my frind again... what he remebers is something like a

5 inch x 5 inch ----- CMYK ----- 300 dpi ----- one Flat colour

took about
- 40 sec - on the -iMac G3- that had only -256 RAM-

and about

- 3 - on his -dual 833 G4- with -768 Ram-

We decided to test this again... on My new machine and his.

I'll post the results shortly

ASUS P4S8X - P4 2.4B - 2 x 512M DDR333 - ATI 9500 Pro(Sapphire) - WD 80G HD (8M Buffer) - SAMSUNG SV0844D 8G HD - LG 16X DVD - Yamaha F1 CDRW - Iomega Zip 250 int.

coylter · Sep 17, 2003

hahaha
i got a
athlon 2700+
512mb ddr400 (single)
2xraptor 32gig hard disk 10000rpm
radeon 9800pro 128mb
and all other thing includint 17" lg flatron f700b
and sub woofer for........
1350 CAN so that mean ...... 900-1000 us
I can have tree computer like mine for the price of the g5 :/

just get a company name and buy from distributor damn it...
you save 20-50% of the price for 20$/year :/

cdpage · Sep 17, 2003

Here are the Photoshop Gaussian blur and @ 250. Pixel Radious results

cdpage
P4 2.4B
2 x 512M DDR333
ATI 9500 Pro
WD 80G HD (8M Buffer)

Jay
G4 833 MP
512 + 256 SDRAM 133
80 GB HD

- TEST ONE - "you can do this at home"

Make a NEW file 50" x 50" 300dpi CMYK and fill it with any colour. Set set the gaussian blur to 250 radius and click ok. (for better results leave the preview OFF)

 2:47 average
 2:55 average

- TEST TWO - "you can do this at home too but w/o the same pic the results may differ"

a file 25" x 25" 300dpi CMYK with an image (we used the same image). Gaussian blur to 250 radius again.

 1:25 average
 1:23 average

ASUS P4S8X - P4 2.4B - 2 x 512M DDR333 - ATI 9500 Pro(Sapphire) - WD 80G HD (8M Buffer) - SAMSUNG SV0844D 8G HD - LG 16X DVD - Yamaha F1 CDRW - Iomega Zip 250 int.
Edited by cdpage on 09/17/03 00:01 AM.

cdpage · Sep 17, 2003

Not exactly the results i had expected...At all.

i thought they would be close with my machine ahead, but thats basicly the same.

SOOOO as for that on PAPER deal My machine is Far better then his;

2.4 Ghz vs. 2 x 833 Mhz
80 8MG Cache vs. 80GB HD ultra ata/66
Radeon 9500 Pro vs. Gforce MX 2 64Mg 
1G DDRAM vs. 768 PC 133 SDRAM

My machine is in every respect a better amchine...on paper.

ASUS P4S8X - P4 2.4B - 2 x 512M DDR333 - ATI 9500 Pro(Sapphire) - WD 80G HD (8M Buffer) - SAMSUNG SV0844D 8G HD - LG 16X DVD - Yamaha F1 CDRW - Iomega Zip 250 int.
Edited by cdpage on 09/17/03 00:35 AM.

imgod2u · Sep 17, 2003

To be honest, you should probably try something a bit more complex than just filling it in with a solid color and running a filter. That kind of data arrangement makes an almost perfect SMP setup (2x 833Mhz WOULD equal 1x 1666 MHz).

Try the test with an actual high resolution photo. The results should be quite different. And probably match real-world situations more.

And yes, a 1.666 GHz PPC G4 (non-plus) would probably be as fast as a 2.4B P4 in this kinda stuff.

"We are Microsoft, resistance is futile." - Bill Gates, 2015.

cdpage · Sep 17, 2003

Well yes thats what the second test was about...we used the same photo.

i just didn't think that the G4 a 1.66 would be faster then my 2.4 considering the 4 reasons i listed that my machine would be faster.

thought it'd be close but still.

I use both Mac & PC and am not a fanboy of eather platform...honestly they both suck, but thats for another discution.

I am glad that the G4 beat mine... to some extent. hopefully some of the "PC happy" people here might take notice of this, Yes, ONE simple test.

ASUS P4S8X - P4 2.4B - 2 x 512M DDR333 - ATI 9500 Pro(Sapphire) - WD 80G HD (8M Buffer) - SAMSUNG SV0844D 8G HD - LG 16X DVD - Yamaha F1 CDRW - Iomega Zip 250 int.

Match these G5 Specs

Distinguished

Distinguished

Polypheme

Distinguished

Polypheme

Distinguished

Distinguished

Distinguished

Splendid

Distinguished

Splendid

Splendid

Splendid

Splendid

Splendid

Distinguished

Distinguished

Distinguished

Distinguished

Distinguished

Distinguished

Distinguished

Distinguished

Distinguished

Distinguished

Share this page