Quick Jump Daily Digest

Thank you for your interest in the Quick Jump Daily Digest. Get notified of all new content on QJ in our free Daily Digest. To subscribe, enter your email address below and click the subscribe button.

Cancel and Return to page

Geeks record PS3's A-OK performance

Posted Nov 23, 2006 at 1:00AM EST by QJ Staff

Listed in: PS3 Tags: Fedora, G5, Linux, MAc G5, Mac OS X, playstation 3 updates

Are you having sleepless nights already playing Resistance? Or are you still trying to install Fedora Core 5? We know you can't get your hands off your new PlayStation 3 and so the guys over at online site GeekPatrol. Being the, well, geeks that they are, GeekPatrol wasn't content with just playing or exploring the PSP-like XMB of the console.

In case you're wondering what are they up to, be informed that they just managed to get a beta version Geekbench for Linux PPC and compared PS3's performance against one of the first Power Mac G5. The setup is as follows:

PlayStation 3: Cell Broadband Engine @ 3.2GHz, 256 MB RAM, Fedora Core 5, Geekbench 2006 (Build 243)
Power Mac G5: PowerPC G5 @ 1.6GHz, 1280 MB RAM, Fedora Core 4, Geekbench 2006 (Build 243)

A PowerMac G5 1.6GHz running Mac OS X was used as the metrics for the experiment, meaning that setup will yield a score of 100 when measured. Power Mac G5 got a score of 106.9 while the PS3 got 105.2. We are then suddenly reminded of earlier comments that the PS3 is actually a stripped down PowerPC-based like current Apple chips. Judging from the numbers, it seems that the console is indeed A-OK. Anyway, here are the rest of the results of the experiment:

Integer Performance

Emulate 6502 (single-threaded scalar)

PlayStation 3: 42.1

Power Mac G5: 73.9

Emulate 6502 (multi-threaded scalar)

PlayStation 3: 57.3

Power Mac G5: 73.8

Blowfish (single-threaded scalar)

PlayStation 3: 118.7

Power Mac G5: 107.0

Blowfish (multi-threaded scalar)

PlayStation 3: 165.6

Power Mac G5: 107.0

bzip2 Compress (single-threaded scalar)

PlayStation 3: 89.8

Power Mac G5: 162.8

bzip2 Compress (multi-threaded scalar)

PlayStation 3: 124.1

Power Mac G5: 168.4

bzip2 Decompress (single-threaded scalar)

PlayStation 3: 76.6

Power Mac G5: 129.9

bzip2 Decompress (multi-threaded scalar)

PlayStation 3: 99.5

Power Mac G5: 133.1

Floating Point Performance

Mandelbrot (single-threaded scalar)

PlayStation 3: 49.0

Power Mac G5: 100.0

Mandelbrot (multi-threaded scalar)

PlayStation 3: 72.1

Power Mac G5: 100.0

Dot Product (single-threaded scalar)

PlayStation 3: 120.0

Power Mac G5: 100.8

Dot Product (multi-threaded scalar)

PlayStation 3: 119.3

Power Mac G5: 100.3

JPEG Compress (single-threaded scalar)

PlayStation 3: 70.7

Power Mac G5: 103.0

JPEG Compress (multi-threaded scalar)

PlayStation 3: 94.8

Power Mac G5: 103.0

JPEG Decompress (single-threaded scalar)

PlayStation 3: 61.6

Power Mac G5: 119.0

JPEG Decompress (multi-threaded scalar)

PlayStation 3: 72.9

Power Mac G5: 119.2

Memory Performance

Read Sequential (single-threaded scalar)

PlayStation 3: 51.9

Power Mac G5: 116.7

Read Sequential (multi-threaded scalar)

PlayStation 3: 56.9

Power Mac G5: 116.0

Write Sequential (single-threaded scalar)

PlayStation 3: 194.6

Power Mac G5: 104.7

Write Sequential (multi-threaded scalar)

PlayStation 3: 191.4

Power Mac G5: 112.7

Stdlib Allocate (single-threaded scalar)

PlayStation 3: 43.4

Power Mac G5: 56.4

Stdlib Allocate (multi-threaded scalar)

PlayStation 3: 51.2

Power Mac G5: 55.6

Stdlib Write (single-threaded scalar)

PlayStation 3: 331.5

Power Mac G5: 92.7

Stdlib Write (multi-threaded scalar)

PlayStation 3: 365.9

Power Mac G5: 94.7

Stdlib Copy (single-threaded scalar)

PlayStation 3: 64.5

Power Mac G5: 63.5

Stdlib Copy (multi-threaded scalar)

PlayStation 3: 102.1

Power Mac G5: 72.7

Stream Performance

Stream Copy (single-threaded scalar)

PlayStation 3: 89.7

Power Mac G5: 114.1

Stream Copy (multi-threaded scalar)

PlayStation 3: 109.9

Power Mac G5: 111.8

Stream Scale (single-threaded scalar)

PlayStation 3: 69.2

Power Mac G5: 118.3

Stream Scale (multi-threaded scalar)

PlayStation 3: 101.4

Power Mac G5: 120.1

Stream Add (single-threaded scalar)

PlayStation 3: 62.6

Power Mac G5: 123.0

Stream Add (multi-threaded scalar)

PlayStation 3: 93.2

Power Mac G5: 118.0

Stream Triad (single-threaded scalar)

PlayStation 3: 62.7

Power Mac G5: 122.8

Stream Triad (multi-threaded scalar)

PlayStation 3: 102.2

Power Mac G5: 118.6

Via GeekPatrol

This story sucks? This story rocks!

Vote Now!

(0)

(0)

Turns out that the benchmark program used is unable to utilize the eight Cell engines. I would expect the numbers to look radically different if the program was able to actually utilize the whole CPU. On the other hand, the benchmark wasn't able to utilize the Altivec features of the G5 either, so...

# yep — Spector 2006-11-23 05:16

I read the same thing elsewhere. So this benchmark means absolutely nothing.

# cell — Guest 2006-11-23 06:27

Yeah this test does not use any of the 7/8 SPUs. Doesn't the Cell also have Altivec?

# The Cell is NO 8core processor! — Guest 2006-11-23 07:25

and they already used the SPUs! Ever heared of a thing called "multi-threaded"?

On a real 8core processor it would run alot faster!

BUT

The Cell is only a single core CPU with some cripled 8kb cached SPU's that can't do anything without the CPU's commands.

When it comes to numbercrunching and stuff like this... the Cell is flawless!

But tell me a game that uses an engine that needs to crunch numbers instead of drawing graphics!

Even a pure physics demo would run faster on a "core2duo" from intel for example!

# The cell is NO 8core processor — Guest 2006-11-23 08:38

Q: does threading clean out the cache every context switch?

Threading has nothing to do with the data-oriented design of

tasks within 256kb caches per SPU operated indepently using

the DMA machinery- a major component not mentioned much.

The PPC is not fantastic its not ment to be. Please read IBM's

reasons for the rejection of level1/level 2 caches in multi-cored

processors to data-oriented stream processors.

Such benchmarks are completely dishonest when they run the CPU as a single ALU for goodness sake it does not have the 2MB

of main cache because it has the SPUs et al.

# of course this only benchmarks the PPC core! — Planet 2006-11-23 09:07

And i'd say it is rather impressive! The code is not adapted to Cell at all, not even optimal for the In-Order-Execution-only PPC core. The SPUs sleep in boredome meanwhile, yet the benchmark runs quite ok.

That only means one thing: while the PS3 surely is no turbo beast in Linux with unmodified programms, it will run them well enough (lack of RAM hurts alot more). Much more Performance would be achievable with:

- compiling the code with a Cell specific compiler

- optimizing the code for In-Order-Execution

- spreading the execution on the SPUs (even with with unoptimzed code they could still do something)

# CELL processor amazing — Guest 2006-11-23 09:15

What's amazing are the university top end high parallel computing

research centres that are very keen on the CELL at the moment.

Homebrew will be a revolution because there will be so much

open source code output by these people.

# i think — Guest 2006-11-23 09:33

I am pretty sure that this isnt that great of performance. Because these guys tested a g5 against a core duo (not a core 2 duo), and the core duo had around a 50% point gain in multiple thread performance. and most of the times it was 2x the points of the g5. and a couple of times it was around 3x the performance. granted that g5 was a 1.8 ghz, and the core duo was a 2ghz. now with the core 2 duos being even faster, the ps3 is a sub-par computer... and dont kid your selves, no one will bother re-writing applications for multiple cpu's, so this is the real life performance that you will expect to get with your ps3... sure there might be one or 2 applications that will have multiple core handling, but 99.99% of all applications will utilize what this benchmarking tool was using.

# the sad part is... — Guest 2006-11-23 09:37

the g5 they tested was a 1.6ghz, so the core duo's performance in real life will CRUSH the ps3's

# Omg — [BOT] Bob 2006-11-23 19:55

Do you even know how a computer freaking computes graphics? It uses numbers and does some crunching then the pixels just magically transform into the picture that is on the back of the game box. Yeah and remember Intel's wonderful HyperThreading, that was on single core processors before they were called single core... and most processors have the ability to alternate between two tasks(yes even each individual core on any current processor, but not the SPUs but more on that later) so while one is waiting for the cache to refresh/fetch data the other task gets to use the core's pipeline. Although current out of order processing is much less cumbersome than Intel's original iteration. Plus out of the nine (thats 9 in case you can't spell numbers) cores on the cell only one, the PPU, has the POWER instruction set, therefore; a linux distro that runs on ppc architecture will only run on that core. Whew and talk about chaching problems, thankfully the SPU's have a 256kb local store registry thingymajig because they can feed the bandwidth hungry SPUs ten times faster than any current cache. Anyways the SPUs are simplified processors yes, but back in the mid 70s some dude figured out that the more you add logic to a system each consecutive step has less of a performance gain, essentially in the end it would be better to remove logic and just make the darn thing go faster(the new intels are toping out at 2.6/2.8 Ghz same with the AMDs) notice that the cell is 3.2Ghz in the PS3(4.0Ghz in more expensive stuff). By the way physics IS number crunching so you contradicted your retarded self. I've read at least five articles about how the cell destroys every other processor in scientific applications(wh ich in case you didn't figure out uses a little bit of physics(oh sorry I was being sarcastic I meant A LOT(just didn't want you to miss that))) by monstrously magnificent magnitudes which would make you feel even more dumb about writing that last sentence; in fact, I would be posting links to those articles right now but I'll just say all of this stuff like its true--kinda like you did--plus you're not worth it.

# if.. — Guest 2006-11-23 20:17

If.... its compiled for a dual core processor, like this program is compiled for a single cored processor, against sony's 8

# . — KirbySS 2006-11-24 01:30

The Cell isn't multi-cored.

# CAN CRUSHING — Guest 2006-11-25 13:34

most code whether it runs on 0.5, 1.5 or 4 GHz will probably

run exactly the same wall speed (what you experience).

This is because the code probably is not written to fit within

the confines of a 2MB/1MB cache. The so called fast ALU just

idles. Now with a GPU a large part of the cache will be taken

by a pipe to the GPU.

Code written for parallel SPU (a lot of data-oriented redesign here)

cannot BUT run unhidered hence the GFLOPs whatever stated

may indeed be representitve of actual experienced performance.

# Power Mac G5 — Power Mac G5 2010-11-25 01:21

Apple simplicity lets you add a Power Mac G5 to your creative studio without days of downtime configuring a new system. The new Mac OS X setup assistant helps you effortlessly move user accounts, system preferences, documents and applications from an old Mac to a new Power Mac G5 and the transfer is FireWire fast.
http://www.applereviewed.net

Refresh comments list

Add comment

JComments

HOT GAMING NEWS | LATEST GAMING SEARCHES | FORUMS

QUICK JUMP DAILY DIGEST

Welcome to QJ.Net!

If you want your comments to go live without waiting for moderation, you need to be logged in. Being logged in has its benefits:

Logged in members do not wait for their comments to be approved.
Logged in members can create Profiles to be seen by other users.

So why wait? Create an account or login now! It's easy, quick, and free.

To get started, use the LOGIN boxes, or the REGISTER link below!

Want to learn more about the team who brings you the QJ news?

Read about them now!

HOT FLASH GAMES

QJ.NET Categories
• Apple • E3 • Featured Articles • Gadgets • Games for Windows • iPad • iPhone • MMORPG • Mobile • News • Nintendo 3DS • Nintendo DS • Opinions & Analysis	• Photography • PlayStation 3 • PoV • PS Vita • PSP • Rumors • Science • Site News • Tabula Rasa • Wii • Wii U • World of Warcraft • Xbox 360

Sitemap

Newest Downloads

05/22/2012 - 05/29/2012

edicpopDS v0.3.01	05/29/2012
RGLoader 0v290 Beta	05/27/2012
WiiMod 3.2	05/27/2012
WiiMC v1.2.7	05/25/2012
True Blue Firmeware v2.7	05/24/2012
3DS Video Converter v2.3	05/24/2012
Mobile Assault 1.8.2	05/24/2012
PSP Tools Alpha 0.1	05/22/2012
Newton War	05/22/2012
DaedalusX64 r766	05/22/2012
PSPMarble	05/22/2012
The Undead 4 - Nazi Zombies	05/21/2012
3DSExplorer v1.2.1	05/21/2012
J-Runner v0.2 Beta 275	05/19/2012
Priiloader 0.8 Beta 3	05/19/2012

Most Comments this Month
Sound Off: On Game Expirat..	(58)
Say Goodbye to the Xbox 36..	(31)
EA Blames iOS Rock Band Ex..	(28)
Flash0 Contents Of Vita's..	(27)
Sony Had a Rough Year; Vit..	(26)
PlayStation 3: Now Featuri..	(23)
Sound Off: Toughest Game C..	(20)
Apparently You Can Pre-Ord..	(19)
Love It or Hate It, It's A..	(16)
Pre-Order Generic Shooter..	(15)
PoV: This Collectors Editi..	(15)
You've Been Up in Arms Abo..	(15)
Get Off Xbox Live, You Ped..	(14)
This is Next Gen?	(14)
Metal Gear Solid For Vita..	(14)
PoV: Paid MMOs are Dead	(14)
PoV: You Are Wrong about C..	(14)
Stop Your Whining, Start C..	(13)
Does Microsoft Ever Have a..	(12)
Operation Raccoon City Pro..	(12)
Diablo III Getting Mauled..	(12)
Pokemon Liquid Crystal! K..	(12)
OK, So Maybe Microsoft Can..	(12)
Update: EA Drops Free BF3..	(11)
Max Payne 3 Will Eat Your..	(11)

Top Jumps this Month
Naruto Shippuden: Ul..	(166)
Pokemon Liquid Cryst..	(101)
neur0n Knocks Out An..	(100)
Bust Out Your PSP Fo..	(90)
PSP homebrew - LameC..	(67)
Wii homebrew - WAD M..	(52)
PSP Comics Reader v2..	(50)
PSP Homebrew: 6.xx P..	(43)
Hellcat Working On R..	(41)
PSP Homebrew - PSPDi..	(41)
PSP homebrew - DSonP..	(38)
neur0n Gives You Mor..	(37)
Sound Off: On Game E..	(35)
PSP GBA v1.2: Gamebo..	(33)
Latest Left 4 Dead 2..	(31)
Another CFW Goodie F..	(30)
GTA IV patch 1.0.6.0..	(29)
Soywiz Releases PSP..	(29)
Download full Sid Me..	(28)
Sound Off: Should Ap..	(28)
Homebrew - PSP/PS3 P..	(24)
PSP homebrew - Mario..	(23)
DS homebrew game - S..	(21)
Brian Lara Internati..	(21)
Metal Gear Solid For..	(21)

HOT GAMING NEWS TOPICS

Quick Jump Daily Digest

Geeks record PS3's A-OK performance

Related Stories:

Comments

Add comment

Newest Downloads