W510 Faster with 4 SODIMMS than 2

I haven't seen mention of this anywhere but my Win7 WEI memory ops per second score went from 5.9 to 7.1 when I went from 2 to 4 SODIMMS - quite a big improvement.

I don't think that just increasing the amount of ram would improve this score but I'm not sure.

Can anyone confirm that the QM57 chipset can interleave the extra modules?

 

Don't rely on WEI scores. You could run benchmarks like 3Dmark06, Vantage, etc and see if you get a higher score with 4 DIMMS.

 

Yes, WEI scores are meaningless and doesn't accurately reflect performance in general.

 

Actually, I'd say that's exactly what WEI scores do do... reflect performance in general.

They do it accurately too - it is a pretty consistent number.

I've yet to see real world usage of a 6.9 machine be whooped by a 1.0 machine, which is what meaningless would imply. It is a set of numbers to use as a comparison, which is all any benchmark is.

Is it as comprehensive as other benchmarks? No, of course not. Can you see if memory throughput increased? No. Does this matter to the average user?

In this instance, more memory was added and the WEI memory performance went up. That is a fair comparison and reflects the advantage of having more memory in the machine. If your machine is memory constrained, adding more memory will increase the performance, as shown by the increase in the WEI score here.

Benchmarks as a whole are pretty meaningless as they test use of a computer in one way, and users use it potentially in another, however for a general performance comparison, WEI is actually pretty good.

 

Nonsense. WEI is the most useless performance analyser. My W510 got higher marks than my M17X in everything including graphics and the M17X tears the ThinkPad apart in every possible bench: 3DMark06, Vantage, SuperPi, WPrime, any and every game, boot and shutdown times, u name it!

WEI is relative to the system. It doesn't really mean much. Windows doesn't even know how to compare 4 GB of 1066 RAM vs 2GB of 1333
Run some real tests if need to know the approximate performance.

 

@realwarder, I'm with you on this one.

I've confirmed it with a 'real' benchmark using Everest 5.50.2100:

# of modules	Read MB/s	Write MB/s
2	8190	8250
4	11050	9015
	(+35%)	(+9%)

The figures are averaged over 20 tests - Everest gives figures with a few percent spread.

I'd say that read figure is definitely statistically valid...

Makes me really happy I went for the W510 over the T510!

 

Very interesting results!
So the system does run better with 4 DIMMS.
That's something to ThinkPad about
+1 rep for the finding!

 

I have seen machines with a score of 4-5 (older but more powerful GPU) beating machines with a score of 6+ in terms of GPU FPS. They aren't that consistent as exact same configurations can have different scores (business orders of several T400's for example yielded slightly different WEI). Generally, I say WEI is useless since it doesn't reflect anything. 3GB DDR3-1067 can give a higher score than 4GB DDR2-800. Is less (but faster) memory better? And by how much? WEI doesn't answer this...

Statistics can be valid but that doesn't mean it's significant in reality. You can't take statistics alone, but you have to look at the big picture. Memory bandwidth scores can be 100% higher, but will you notice a difference in any real life situation? No, you won't.

 

WEI does reflect SOME aspect of system performance but please understand that its testing is limited in comparison to extensive benchmarking apps.

Benchmarks often don't matter as much in real life, so WEI is even less significant often times. That is not to say it does not always provide meaningless results.

 

I think we're getting a bit OT here...

The thread is about an increase in speed going from 2 to 4 dimms - I wouldn't disagree that WEI is a low grade benchmark or that you probably wouldn't notice the difference editing a word doc but you can't really argue with a 35% boost!

 

"WEI does reflect SOME aspect of system performance but please understand that its testing is limited in comparison to extensive benchmarking apps.

Benchmarks often don't matter as much in real life, so WEI is even less significant often times. That is not to say it does not always provide meaningless results"

Jaredy,
Usually Im in agreement with you and then I got to the second paragraph. A benchmark is as good as it's author. A well written benchmark is applicable.
Renee

 

I just populated all 4 memory sockets with 2GB sticks. Everything seem to run faster and the memory WEI jumped from 5.9 to 7.3. I only had 4GB of memory installed originally. What a difference it made !

 

Are all SoDimms the same brand/size?

eg. 4 x 1GB DDR-3 Kingston xxxxMHz or did you mix sizes, brand and stuff? Basicly all slots should work with dual channel mode (triple channel is only available on desktop machines I think) so more memory working in dual channel mode should increase the Windows 7 benchmark results.

 

More memory (above 4GB) will not provide any noticeable benefit in any real life basic task (Firefox, Word, Excel, etc.), though specific programs such as MATLAB and Photoshop can take advantage of it. It also won't make a big difference if the SODIMMs are of different capacity or speed since the chipset will run asynchronously.

 

Mine are different brands, sizes and speeds (2 x 2GB PC3-8500 + 2 x 4GB PC3-10600) but the chipset still seems to interleave them.

The point of this thread is not about the quantity of memory, it's about the increased throughput by having 4 modules installed - they could be 256M each.

This also won't make much difference to word or firefox but neither will most upgrades!

You'd think this would be the case but it isn't. See above.

 

the W510 can support ram pair running at different speeds.