Just got my X60, got some early benchmarks for you

I will be writing up a proper review, but here are some benchmarks for now ...

This is the top X60 model with a T2500, 2GB of 533MHz RAM, etc. I am expecting 2GB of 667MHz RAM to arrive soon, which is the configuration that I will be benchmarking.

Video card speeds


Super PI


HD Tune
[/IMG]

3DMark05


3DMark06

 

Welcome to theSamsung X60 owner's club.

John

 

Thank you, I love it so far. My ONLY complain is the stupid FN <--> CTRL situation ...

 

Wow, these performance data is pretty nice ...

 

And they will only get better as I just received my 667MHz memory modules and I am running the benchmarks again

 

I have just upgraded the RAM of my X60 to 2GB. Measurements of performance are summarised below:

Using 1GB of 533MHz RAM (Transcend)
Memory timing 4-4-4-11
Sisoftware Sandra memory benchmark:
RAM Bandwidth Int Buff'd iSSE2:3354 MB/s
RAM Bandwidth Float Buff'd iSS:3342 MB/s
PC Mark 05: 3468


Using 1GB of 667MHz RAM (Corsair)
Memory timing 5-5-5-15
Sisoftware Sandra memory benchmark:
RAM Bandwidth Int Buff'd iSSE2:3502 MB/s
RAM Bandwidth Float Buff'd iSS: 3497 MB/s
PC Mark 05: 3454

Using 2GB (1GB of 533MHz RAM (Transcend) + 667MHz RAM (Corsair))
Memory timing 4-4-4-12
Sisoftware Sandra memory benchmark:
RAM Bandwidth Int Buff'd iSSE2:3493 MB/s
RAM Bandwidth Float Buff'd iSS: 3616 MB/s
PC Mark 05: 3445

The results are difficult to interpret. 2GB gives the lowest PC Mark 05 score, but only by a small fraction. (I only ran the tests once). The single / dual channel issue had no effect.

Sandra's results suggest that the faster clock speed of 667MHz RAM is offset by the longer timings.

For me, the real objective of adding more RAM was to reduce the use of the swap file - often I have many applications open.

John

 

That's interesting. For optimal performance it is recommended that you use exactly the same memory modules for both slots, but I don't know exactly what the side-effects are when you use memory modules that have different specs.

Have ended up with 2GB of 533MHz and 5-5-5-15 timings because the system needs to apply the same settings on both modules?

 

Ideally, there should be matched modules. If not, the system will use the speed of the slowest one. In my case I seem to have ended up with 533MHz and the 4-4-4-12 timing. It appears, from the results, that 667MHz and 5-5-5-15 timing is no faster. The timing units are one clock cycles, so 5 clock steps at 667MHz is the same time as 4 clock steps at 533MHz.

The other factor in the overall performance could be the timing difference between the CPU and the RAM. In theory, 667MHz should be better matched to the Core Duo, but in reality there seems to be no difference in performance.

Maybe there will be a significant difference in RAM performance when 667MHz moduleswith CL=4 are used. At present these are hard to find.

John