Thinkpad W520 PCMarkVantage Benchmarks - Post Yours Here!

Here are some PCMarkVantage Is Your PC Fast Enough? PCMark Vantage Basic Tells You | PCWorld benchmarks:

W520 i7-2820 8GB 1333 Mhz RAM ( 500GB HDD): 8,614 PCMarks
W520 i7- 2920 8GB 1333 Mhz RAM (500GB HDD): 9,970 PCMarks

W520 i7-2820 16GB 1333 Mhz RAM (single OCZ Vertex3 SSD): 16,006 PCMarks
W520 i7-2820 16GB 1333 Mhz RAM (single Intel 510 SSD): 16,302 PCMarks
W520 i7- 2920 16GB 1333 Mhz RAM (single OCZ Vertex3 SSD): 18474 PCMarks

W520 i7-2820 8GB 1333 Mhz RAM (2x240GB Intel 510 SSD in RAID0 - 128k stripe): 19,000 PCMarks
W520 i7-2820 16GB 1866 Mhz RAM (2x240GB Intel 510 SSD in RAID0 - 128k stripe): 19,228 PcMarks
W520 i7-2920 8GB 1333 Mhz RAM (2x240GB Intel 510 SSD in RAID0 - 64k stripe): 20,700 PcMarks

Results:
i7-2820 vs i7-2920 improvement: +15%
HyperX improvement: +1%.
Sata III SSD improvement: +89%
RAID0 improvement: +16%

Your results may vary. Please post yours, especially if they are better. If you have a different model or brand it would be interesting to see improvements on these results.

 

Discrete Graphics W520 i7-2720 16GB 1333 Mhz RAM (single 256GB Crucial M4 SSD): 14,545 PCMarks

Integrated Graphics W520 i7-2720 16GB 1333 Mhz RAM (single 256GB Crucial M4 SSD): 13,004 PCMarks
*Forgot to switch it before I ran the test so I threw it up here for comparison sake*

 

tried again after wiping and changing fw on ssd and changing nvidia drivers, here it is.
Result

 

Did you have the "write back" cache on? (programs>intel>rst>manage)

 

18,834 !!!. Whew! That i7-2920 rocks! (FirmWare update helped too.) Your results and those of storagereview.com convinced me to exchange my W520 i7-2820 for an i7-2920!!!

 

You know the disk benchmarks are designed for platter drives and SSDs totally screw up the value of the point system, right?

 

Hmm, I have never heard that. Do you have link you can refer me to?

As far as I understand, PCMarkVantage measures a series of typical productivity tasks. It just happens that SSDs are helping the system to get them done faster...

 

My W520, I7 2820QM, C300 128 GB OS, C300 256 GB Data Drive, quadro 1000 got 9448 PCMarks while my X220, I7 2620, Intel X-25M, got 12079 PCMarks?!

I am much happier with the old Gen 2 SSDs than the new Gen 3s. That is the only reason I can see that the quad core with quadro would test slower than a dual core with less graphics power.

Yup, in checking the detailed results all the HDD intensive tasks in the benchmarks are much quicker on my Intel X25-M than on my Sata 3 C300s and all the tweaks on each system are identical. I just think the older Intel controllers on the larger Nand media work better with the current machines.

Interesting enough the SSDs all benchmark about the same with AS SSD or CDM.

 

Think about it: Platter drives and SSDs operate in completely different ways and bear very little resemblence to each other. A benchmark valid for judging the performance of platter drives will generally not hold for SSDs.

Is increase of 2000 points for the disk score from 16,000 to 18,000 really worth the same as from 2000 to 4000?

For example, if you RAID 0'd your 500GB platter drives, would the performance delta between the non-RAID setup be worth the increase in the synthetic benchmark?

 

I would have to disagree. The real world tasks run by PCMark Vantage don't care what the hardware is they just need to occur and be timed.

If an SSD gives a better result than a platter HDD that is because it is faster.

Just like I am willing to pay more for a SSD than a HDD because it is naturally faster, not because it can benchmark a higher number. The benchmark is only the measure. The result we want is the speed. So if the benchmark is at all valid it should give a better result to a SSD equipped computer.

 

Sure, SSDs are generally faster. But is the real-world performance worth that many points in the benchmark?

 

Yeah, that's why you have to look at the sub-scores of the different tasks and applications and pick the one that is most important to you.

Remember, PCMarkVantage is not a synthetic test. It uses typical productivity tasks.

It is the best benchmark I could find that approximates what I do. If I played games, I would use the 3DMark benchmarks, for example.

 

Err... It is a synthetic test. These tools all count as synthetic benchmarks. A real-world benchmark is actually doing the productivity task in question.

See:
Techware Labs - Articles - Synthetic VS Real World Benchmarks
Synthetic Benchmarks: 3DMark And PCMark Vantage : How Many CPU Cores Do You Need?
etc.


Consider the following data:
http://techreport.com/r.x/scorpio-black-750/hdtune-read.gif
The top of the list, RAIDed X25-V's, would no doubt have extremely high benchmark scores. Likely, you'd see 5-7x the score value.

Then consider this:
http://techreport.com/r.x/scorpio-black-750/time-boot.gif
The same configuration leads to a few second decrease in boot time.

Then there's this:
http://techreport.com/r.x/scorpio-black-750/time-load-crysis.gif
Crysis doesn't load 5 to 7 times faster, does it? Not even twice as much.

This is the benchmark score for a Firefox compile:
http://techreport.com/r.x/scorpio-black-750/db-compile.gif
This is an actual Firefox compile:
http://techreport.com/r.x/scorpio-black-750/firefox-compile.gif

 

This is a synthetic tests that only test I/O. A Firefox compile is more dependent on CPU than on I/O, so the results are not surprising.

PCMark Vantage shifts its focus from only providing pure hardware performance-analysis results, to generating accurate performance ratings based on popular Windows Consumer Scenarios ranging from media and entertainment to productivity and communication.
|MG| PCMark Vantage

 

i can tell you one thing from a real world perspective. the ssd's make the w520 seem like a totally different machine. i bought the vertex 3 240gb when i ordered the w520 and it came in weeks before the w520 so i tried it in my t61p. i almost canceled the order for the w520 it was so much faster. from first boot on. with the v3 i can go from a cold machine to a rendered web page in around 50 seconds and that's with a windows password. everything else seems that much faster as well.

 

I don't know how much fast SSD matters to T61p because of it's SATA cap. It's definitely much better than mechanical drive...

- ngm

 

I don't even think I have Intel Rapid Storage Tech installed. I don't have an intel ssd. Should I download and install it anyway?

 

Yes, it is not just for Intel SSDs.

 

Installed the version for Series 6 chipset and it is working fine, however, under Manage there are no changeable options. Just a list of information.

 

programs>intel>rst>manage>advanced

 

You definitely need to google for SSD tweaks and read those threads.

The write change can be made in windows without any help from Intel.

There are a number of other changes you might want to make, but write caching is a certain one.

I find the Intel controller drivers a good thing on my Intel X25-M. However on my Crucial Gen 3 drives I am definitely better off with the MSAHCI drivers. They are much more stable.

 

Before I did anything else, I went through this guide and did most of them that applied.

The SSD Optimization Guide | The SSD Review

As you see for #8, they do not really recommend turning off write-cache, so I did not do it.

 

For some reason I can't make PcMark run with the nVidia GPU by right clicking the icon and choosing the GPU. I have to manually change it in the BIOS. Is this normal?

Ok, I just figured out how to add it in the "Manage 3D settings". Sorry for being a noob.

Edit 2: Well, some of the test ran with the nVidia, but not the whole thing. Might just have to switch it manually every time.

 

OK, you have write cache enabled.

 

Guys, I re-ran the benchmark (multiple times) with write-cache disabled and my scores dropped 2,000. I re-enabled it and they went back up. It must be better for Crucial SSDs to leave it enabled.

Write-Cache ENABLED = 14,384 DISABLED = 12,351

 

I am sorry if you have not been understanding us. But you are correct enabled is much faster with any SSD.

I would very much appreciate access to the detailed scores of your W520 with write cache enabled. My W520 with Crucial C300 is sickly slow as previous posts mention. Possible you could screen shot the detailed info or (if this board allows it) screenshot it to me in a private message. I have the comparison between my X220 and my W520 but would like to see a W520 with Crucial C300 for a more direct comparison.

My X220 is over 12,000 but my W520 which should be about 14,000 is 9,000.

 

@huberth and other having RAID enabled motherboard
The first mSATA SSD with 6GB/s is just announced: RunCore outs T50 6Gb/s mSATA SSD, SandForce speeds included -- Engadget. Just imagine, with that one and 2 SSD's in RAID0, what would be the PCMarkVantage benchmark?
@ALL
Now it would be very interesting to discover if that mSATA SSD would expand the whole 6GB/s speed in the W520?

 

Like I wrote in another post

 

I couldn't help myself. this is the same machine that scored the 18474 with the vertex 3 drive. with an intel 510 drive, and a lot of trial and error,the score now is 19,792. this is with a single drive! here's the link.
NVIDIA Quadro 2000M video card benchmark result - Intel Core i7-2920XM Processor,LENOVO 4270CTO score: 19792 PCMarks

 

Whew!

What tweaks did you do to get such a great score?

 

it's hard to remember everything! the biggest change i got was with ramdisk. i tried different sizes between 500mg to 4 gig and got the best result at one gig.i set it up to save at shutdown and load at reboot. i set it up and format as ntfs and start out putting on internet explorer and chrome cache on it. ie cache is set at 250mg. i reboot a few times to make sure everything works. then i move the temp and tmp folders to it. reboot to make sure it's working. then i disable the page file on the main drive and set it up on the ramdisk. i use 250 mg for the size. once i got it all sorted out i imaged the drive. yesterday i tried to increase the ramdisk size to 2gb and got scores back in the 17500 area. i don't disable indexing on the drive but i do disable defrag,system restore and hibernate. i'm sure there are several small changes that i don't recall. most of it's trial and error.

 

Did you use Dataram RAMDisk? RAMDisk - Software - Server Memory Products & Services - Dataram ?

It makes sense to move internet explorer and chrome cache and temp and tmp folders to it.

This should also prolong the life of your SSD

 

i used the dataram. i didn't think moving the page file would help but with mine it did. i tried not having a page file but windows created one every time. when i moved it to memory it seemed to speed things up so i left it there. should be less writes to the ssd as well. i have no idea the scores dropped when i went to a 2gb ramdisk. each increase in size made for a longer boot and shutdown. with 1gb it cost me about 5 seconds but it was more than worth that to pick up the extra speed.

 

Note: The Alienware M18x seems to be the World's Fastest Laptop now:

PCMark 7: 5,824 - a new World Record for Laptops!
Result not found

(too bad the Lenovo BIOS does not let W520 users overclock the i7-2960XM to 4585 MHz)

 

IMO if things like that start to increase some benchmark score, they serve merely as indicators of shortcomings and limitations of the benchmark, and have very little to do with practical performance. Although I can think of even better idea, shame that Windows does not permit moving hibernate.sys to a RAMDisk for instant hibernation

In practice, most memory-hungry and performance-critical applications, from databases to photoshops, would not run any faster if a lot of available RAM is just taken away, and accessed through relatively inefficient RAMDisk. If the machine has a lot of useless RAM, a better idea is to change file system caching options to use memory more aggressively (and recent Windows versions do a decent job by default too). At least that RAM can be reclaimed back, if an application requires it.

 

So does moving your Temporary Internet Folder and temp folders to your non SSD drive make things run faster, or is it solely to decrease wear and tear? I originally moved it off my SSD but then I figured I would want that data to be as fast as possible because it's used so much. And lets face it, by the time I wear out this SSD there will be bigger and better ones on the market and I'll likely get a new system our just upgrade it.

 

I did mine to reduce wear on the ssd but I moved them to my ramdisk and it actually sped things up. I have 16gb or ram and I use 8 of it for the ramdisk. Once I got it like I wanted it I saved it and that's the image I load every time. It can cause problems when installing programs that require a reboot so if i'm doing that I set the ramdisk to save on shutdown before I load the program. I would recommend installing as many of your programs as you can before setting up like this. I've been using it like this for several months and am very satisfied with the way it works.