isnt it lame?

Obsidian

macrumors member
Original poster
Oct 10, 2001
66
0
Victoria, BC, Canada
this is going to be a short post because I dont have the time to write a long one. I am still without a monitor but i stole my dads for the next few hours.

is anyone else fed up with the 133mhz boosts? thats a total of 266mhz for one year. what has intel boosted? like 600-800mhz?

lets get this thread a rockin so people come a knocking!! :D
 

Ensign Paris

macrumors 68000
Nov 4, 2001
1,781
0
Europe
Who cares?

THE MAC RULES!

NOT EVERYTHING IS BASED ON MHZ!

from now on I measure in Gigaflops:

Intel Pentium 4 2.2ghz - ??????gigaflop Let me know!
G4 DP1ghz 15gigaflops
 

Ifeelbloated

macrumors regular
Jan 14, 2002
245
0
some God forsaken place
Don't you wish that there were some finalizing test that really measured the speed of the top systems? I wonder what the Gigaflop rating is for the AMD and Intel processors. Why the hell can't anybody just do that? Forget all the tech specs and just let the machines do an identical task and see how long it takes to do it. I'll bet that all the processors come pretty close in performance.
 

GigaWire

macrumors 6502
Dec 25, 2001
386
0
there's not much in the way of Pentium or Athlon specs measured in gigaFlops, but here are some Supercomputers to give you an idea. :)


The dual 1GHz PowerPC G4 processors — with a combined performance of 15 billion floating point operations per second, or 15 gigaflops.


Some comparisons:

Computer 1:
The system is configured as one front-end system and 64 compute nodes.


Each of the 65 nodes in the system has:

Motherboard with Intel 440LX chipset and 300 MHz Pentium II processor
128 MByte 10-ns SDRAM memory
3.1 GByte Quantum EIDE U-DMA disk
100 Mbit/s ethernet adapter
The machine has an aggregate peak performance of 19.5 GigaFlops.


Computer 2:
Granovsky ran a version of the General Atomic and Molecular Electronic Structure System (GAMESS) ab initio quantum chemistry program on an eight-processor Itanium-based server cluster at NCSA. Performance on an eight-processor Itanium cluster was more than 12 gigaflops.

Computer 3:

The Halo World project will use a Beowulf system to analyze N-body simulation data. This system consists of 59 Intel Pentium-Pro processors (200 MHz clock) for an aggregate system peak performance of 11.8 gigaflops, 7.5 gigabytes of main memory, and 182 gigabytes of internal secondary storage.

Computer 4:
Seattle's Cray Inc. said it provided an SV1 supercomputer to the U.S. Army Space and Missile Defense Command (SMDC) - the Huntsville, Ala.-based hub for American space and national missile defense operations.

SMDC will use the SV1 to run complex simulations in support of its national defense initiatives, according to a release. The 16-processor, Unix-based SV1 system will later be upgraded with SV1ex processors, each with a peak performance of two billion calculations per second (gigaflops). At a total of 32 gigaflops, the system will command several thousand times the computing power of a personal computer equipped with a single Intel Pentium 4 processor.
 

Catfish_Man

macrumors 68030
Sep 13, 2001
2,579
1
Portland, OR
I'm attempting to make...

...a Total Performance Indicator benchmark. This would consist of the average of tests that measured the speed of each part of a computer relative to a base machine. Here's the categories I've come up with:

OpenGL: QuakeIII or Unreal2
Altivec 2D: Photoshop filters
Non-Altivec 2D: ??? (the GIMP maybe, it's not Mac native so it might not use Altivec)
HD Read: Average of copy times to RAM disk
HD Write: Average of copy times from RAM disk
CD Read (CD/DVD tests optional): Copy from CD to RAM disk
CD Write: Burn from RAM disk to CD
DVD Read: see CD read
DVD Write: see CD write
RAM speed: Copy from RAM disk to RAM disk
Encryption: plenty of benchmarks here, but I need suggestions
Altivec Math: Altivec Fractal Carbon Demo (AFCD) w/ Altivec on
Non-Altivec Math: AFCD w/ Altivec off
Movie: Some iMovie test, I'm sure it won't be hard to find one
MP3: Some iTunes test
Any other tests?

So you would run these on a base machine (I'm thinking a new 800MHz iMac), then compare the scores. So if your machine got 140fps in QIII and the iMac got 70 then your comp would score 2 for opengl. Then you average these scores together to get a general picture of how much faster or slower your machine is than the base. (I'm still thinking about whether to include the CD/DVD tests)

ps. GigaFlops is not a good measurement either. PPC processors use smaller operations than x86 so they score higher on GFlop tests even if they get the same performance.
 

Onyxx

macrumors regular
May 5, 2001
152
0
they may use smaller instructions but thats what makes them faster. now al we need is software that is optimized amazingly well for ppc chips (almost all apple software ex. final cut pro, itunes, imovie are pretty well optimized). If photoshop was optimized to the max for a ppc chip it would be a heck of a lot more efficiant. we've got the power, now all we need is the proper stuff to run on it.

gigawire, imagine a couple of those new dp 1ghz g4's turned into an appleseed style network. specs

2 g4's 30 gflops
3 g4's 45 gflops
4 g4's 60 glfops...he he he, and these are just the first crop of apollos
 

Onyxx

macrumors regular
May 5, 2001
152
0
catfishman- you need a raytracing test (aka a 3d test)
i don't remember what the standard test is but it is cross platform.