is this correct - 8600M-GS > 8600M-GT?

this pic is taken from a review of the xps m1530 done by notebookcheck.net. the pics shows the 8600gs scoring higher than the 8600gt. it that true??
http://www.notebookcheck.net/index....> | </a>&md5=e8be7c7934c381f539f6f3bfd06cc761

 

No. The 8600 GT is far more powerful than the GS version. Usually, for Nvidia, the GS version is the lower clocked one ( check 7900GTX > 7900 GS )

 

Handy guide to Nvidia graphics card nomenclature.

Higher number is better.
Later (alphabetically) suffix is better.

Example:
8600 GTS > 8600 GT
8800 GTX > 8800 GTS

 

actually it doesnt. the other one has a t7100 while the xps has a t7250

 

where did the other posts go

 

PCmark05 isn't exactly the best measure of GPU performance.

Try comparing 3dMark scores instead.

 

but both the cpu and gpu are worse in the other model. how could it score more even in pcmark

 

not entirely accurate...
for example,

8800 GT > 8800 GTS

anyway, the nvidia nomenclature is subject to change, one good way to see what is what is to check it out at tomshardwareguide.com

 

i understand that the 8600gt is more powerful than the 8600gs but how do u explain the pic

 

They could've tested it at its out of the box configuration, meaning no reformatting or removal of bloatware. And Dell is a bad offender at that. So much so it could've brought down the score that much.

Also, I think PCmark is more dependent on both CPU and RAM than GPU, if it even affects the score at all. I've seen laptops with integrated graphics get high PCmark scores.

 

Half right. the OLD 8800 GTS < 8800 GT.

However the new 8800 GTS > 8800 GT

new 8800 GTS ~ 8800 GTX with overclocking

 

Pcmark shows an overall score and like any other benchmark you cant just decide on just one result that it is better than the other, like others said 3dmark is the best way to benchmark GPU if u need a benchmarking program. But otherwise it is "just" numbers which do not mean much it is the real world performance that matters most and in that i am very sure that your comp is far more powerfull than any other with t7100 and 8600GS

 

I believe the 8600m GT has DOUBLE the number of stream processors than the 8600m GS. The GS can partially make up the difference with higher clocks, but there is no replacing the extra raw horsepower.

When will Nvidia adopt a new naming system more similar to ATI's new system? JUST NUMBERS. BIGGER IS BETTER. PLEASE!

 

In terms of raw power, yes. In terms of price-to-performance and valued judgments, no.

For a budgeted buyer, finding the right combination of performance for price is the tricky part. It's easy to simply point out the most powerful card, but it'll come at a heavy price premium.

 

Let's take a look at the GPU stats from nvidia.
(Source http://www.nvidia.com/object/geforce_8600M.html)

8600GS
Stream Processors 16
Core Clock (MHz) 600
Shader Clock (MHz) 1200
(memory amounts, clocks and bus are the same provided both use GDDR3)

8600GT
Stream Processors 32
Core Clock (MHz) 475
Shader Clock (MHz) 950
(memory amounts, clocks and bus are the same provided both use GDDR3)

The 8600GT has twice the stream processors at 80% of the clock speed.


The 8600GT should have more than 1.5x the oomph of the 8600GS in sheer power, but that will be mitigated some by the application and various bottlenecks.

Note that one of those bottlenecks happens when the 8600GT has GDDR2 installed and almost halves its memory bandwidth.

The 8600GT should still win out.

Benchmarks like 3dmark have some downsides.

Their numbers and standard tests are useful for comparing systems, but are not great for comparing GPU alone unless similar systems with only GPU differing are used.

why?

3dmark has a processor-only component which heavily weights the score based upon processor-only performance.

Someone with a weaker GPU and a slightly higher CPU will receive a higher 3dmark score than someone with a powerful GPU and slightly weaker CPU... yet the better GPU and weaker CPU will usually destroy the weaker GPU and slightly better CPU in games.

In order to use any benchmark, you must understand what it tests and how it scores and biases its tests.

note, it is also a bad idea to use only ONE benchmark to test anything as "tuning" to a particular popular benchmark has been used all too often as a marketing tactic.