What video card is in your 2503?

[emeraldgirl08]

Thanks for the tests Medessec

If I can make time this weekend I will give my T500 a separate run between the built-in HD3650 and the docked-HD 6570. I also have a T61 w/NVS140 but I am a bit reluctant to include it as it is not a revised GPU model type. For a bit more info could you include the specs of your card? I know it only takes a few keystrokes to find it on the web but for a quick reference while reading the thread it would be handy

Where did you download the Performance Test 7? I only see version 8 on the PassMark website.

[Medessec]

If I can make time this weekend I will give my T500 a separate run between the built-in HD3650 and the docked-HD 6570. I also have a T61 w/NVS140 but I am a bit reluctant to include it as it is not a revised GPU model type. For a bit more info could you include the specs of your card? I know it only takes a few keystrokes to find it on the web but for a quick reference while reading the thread it would be handy

It'd be good to get a reference point of the T500 ATI graphics... especially compared to the T61's NVIDIA. It's always been a strange battle between those, even though the NVS 140M/FX 570M is older, it's NVIDIA, so it has better long-run potential and efficiency.

You don't have to put your T61 in the line of fire... that's why we have mine Mine is one of TuuS' ridiculously rare NOS boards, so any testing can certainly be done on mine with no worry of a One beep-two short beep. If you have a specific game that you want me to try, such as Bioshock Infinite... I can probably work that in, because I own that particular game. It'd just be the trouble of getting it on my Frankenpad. Other games, I'd probably have to seek out somehow, but I'd be willing to try if you tell me what game.

Again, my card is the ASUS Radeon EAH6450 Silent design. Of course, because it's just any other card, the maker/seller neglects to post power consumption info... although some googling does show someone pointing out a 27W peak. The EAH6450 has 1GB GDDR3 VRAM, full 16x length PCI-E interface, heatsink only design(no fan), and 625MHz core clock. Reviews for it are generally favorable, with some people complaining about the heat and high power of the card, but in contrast, low stamina(inability to maintain smooth or high quality gameplay). Of course, no good product is without it's pool of mindless consumers who don't know how to use it.

Where did you download the Performance Test 7? I only see version 8 on the PassMark website.

I bought Version 7 when it was out... awhile ago. Hopefully version 8 doesn't score much differently. I'll download the demo and see if it does... and I can also get some V8 scores to compare, if it is notably different.

[86turbodsl]

Lately i've been kicking around the idea of a 2503 hack. I have several of these docks, and long wondered if it would make sense to upgrade the power supply, perhaps with a small ATX supply. I haven't taken one apart yet, but i'm assuming there's actually some dead space inside the case, that might make a spot for a second power supply. 50w total isn't very much for a performance graphics card really, although with the limited clearance in the case, a larger heat output may be counter productive.

Anyone ever taken one apart?

[emeraldgirl08]

Okay well I did a quick test with my T500 docked w/MSI Radeon HD 6570 (click here for specs) using the same PerformanceTest Version 7 that Medessec used for his results. I also am in school (full-time) and have a job as well so both keep me very busy! I only have preliminary results at present however I will continue to come back and add to my post in the future as I collect more results. BTW my system specs for the T500 are in blue font.

* asterisk indicates the Lenovo Power Manager settings

T500- W7 64-bit, Intel C2D P9500, Intel/ATI HD 3650 switchable graphics, 4gb DDR3 RAM, X-25m 160gb SSD (Main drive; OS), Samsung Spinpoint 500gb (Ultrabay Hard Drive; misc. files)

T500 using Radeon HD6570 through 2503 Dock

• CPU: 1729.8, 2D: 461.7, 3D: 352.3

T500- docked w/HD 6570/Driver from MSI site

• *Adaptive- Balance all parameters- aero off- PCIE moderate power setting*

• CPU: 1753.4, 2D: 457.2, 3D: 335.9

[/b]

• Highest- Balance all parameters- aero off- PCIE moderate power setting

• CPU: 1768.4, 2D: 460.0, 3D: 174.4

• Adaptive- Balance all parameters- aero on- PCIE moderate power setting

• CPU: 1731.0, 2D: 157.6, 3D: 321.8

• Highest- Balance all parameters- aero on- PCIE moderate power setting

• CPU: 1742.5, 2D: 160.7, 3D: 179.2

T500- docked w/HD 6570/Driver from MSI site

• Low- Balance all parameters- aero on- link power state on moderate
  CPU: 551.9, 2D: 76.9, 3D: 329.3
  
  Low- Balance all parameters- aero off- link power state on moderate
  CPU: 565.9, 2D: 157.8, 3D: 343.3
  
  Low- Balance all parameters- aero off- link power state off
  CPU: 569.7, 2D: 158.4, 3D: 341.9
  
  Low- Balance all parameters- aero on- link power state off
  CPU: 552.0, 2D: 77.0, 3D: 340.7

T500 -undocked/ATI 3650

• Adaptive- Balance all parameters- aero on- PCIE moderate power setting

• CPU: 1725.6, 2D: 251.0, 3D: 272.3

• Highest- Balance all parameters- aero on- PCIE moderate power setting

• CPU: 1738.7, 2D: 262.1, 3D: 270.2

• Adaptive- Balance all parameters- aero off- PCIE moderate power setting

• CPU: 1747, 2D: 464.1, 3D: 287.4

• Highest- Balance all parameters- aero off- PCIE moderate power setting

• CPU: 1759.4, 2D: 465.5, 3D: 279.9

T500 using Intel X4500MHD (integrated graphics)- no dock

• CPU: 1735.9, 2D: 309.3, 3D: 185.3

T500- undocked / Intel 4500MHD**

• Adaptive- Balance all parameters- aero on- PCIE moderate power setting

• CPU: 1730.3, 2D: 273.5, 3D: 185.2

• Highest- Balance all parameters- aero on- PCIE moderate power setting

• CPU: 1738.4, 2D: 273, 3D: 182.3

• Adaptive- Balance all parameters- aero off- PCIE moderate power setting

• CPU: 1751, 2D: 459.2, 3D: 199

• Highest- Balance all parameters- aero off- PCIE moderate power setting

• CPU: 1771, 2D: 468.1, 3D: 199.6

** The Intel graphics could not perform the 4X anti-alias test for the graphics test; not sure how Performance Test 7 adjusts for the score; the numbers are most likely inflated as Medessec’s NVS140m should clearly score higher than the Intel 4500MHD. I will post the scores of another Intel 4500MHD device, my X200tablet, as a control with similar Lenovo Power Manager settings as well.

X200tablet –W7-64 bit, L9400 C2D 1.86GHz, 4GB DDR3 RAM, Intel X-25M 80GB SSD

• Adaptive-Balance all parameters- aero on

• CPU: 1284.2, 2D: 218.5, 3D: 185.1

• Highest- balance all parameters- aero on

• CPU: 1295.4, 2D: 214.2, 3D: 185.3

• Adaptive- Balance all parameters- aero off

• CPU: 1310.3, 2D: 375.7, 3D: 198.8

• Highest- Balance all parameters- aero off

• CPU: 1313.9, 2D: 376.9, 3D: 198.8

MSI HD 6570 video out ports

• HDMI with Audio is functional

Preliminary Conclusions

Power may be a factor here. If you take a look at the scores for 'T500- docked w/HD 6570/Driver from MSI site' testing category then notice how when I have Lenovo Power Manager set the CPU speed to 'Highest' followed by 'Adaptive' that there is a difference in 3D scores. Also to further that hypothesis I included a test which basically restricts the CPU speed to 'Low.' I think the CPU frequency is set to below 1 GHz in that setting IIRC. Notice how the graphics scores are still 300+ while the CPU drops to ~500. In contrast when I have Lenovo Power Manager set to 'Highest' the 3D graphics are halved from 'adaptive' and 'low' settings. It seems that allowing the CPU to draw enough power to run at 'Highest' affects the scores of the 3D test results. That seems to indicate (quite saliently) that the balance between power and performance is extremely inversely proportional with this set-up. More CPU draw power results in less 3D performance. Less CPU drawing power results in greater 3D performance. The 2D performance I gather is CPU dependent so it is not too surprising that the 2D remains well above the 3D in 'Highest' power manager settings. I am not too familiar with how much the Power Manager restricts CPU frequencies so if anyone wants to comment about that please do.

The Intel graphics all seemed to produce the same score (minus the 4X anti-alias test). I expected that and I also expected the results to be higher when the power manager is set to 'Highest' as well however it appears that there is not a significant difference between 'Adaptive' and 'Highest' for the Intel gfx tests. Also having Aero either on or off seems to make a difference in my testing also. I am still convinced that if one has an Intel graphics-only T500/T61/T60/etc then there is a benefit with using a dock+video card which is noticeable if comparing the results of the Intel graphics with those from the HD 6570. The scores between the HD 3650 and HD 6570 may seem close however subjectively I am able to see a difference in performance.

I think I may have run into the 2gb limit when docked+using a gfx card in the 2503. There have been a number of times I got a series of endless cycle of reboots upon starting the T500 while docked. At some point fiddling with the 'Display' settings in BIOS seemed to remedy this. I am not sure what my settings are but after messing with them my T500 will finally start. I have not taken out 2gb to test yet. Too busy with preparing for finals and work (terrible during finals but bills need to be paid). I will mess around with the settings at some point in the future and note which setting works without the reboot-cycle.

Finally if I made an obvious typos it may be because my circadian rhythm is FUBAR at this point lol. I need food also right now so I am going to get ready to go out and get some. Need brain food!

• 05/23/2014 Update

Have been using the dock w/HD6570 on-and-off. I have a small break before I become busy with a summer internship so I decided to visit the family and squeeze in a little tech work as well. So my deal with the advanced dock 2503 thus far is that it is terribly noisy. I am sitting here in a somewhat silent environment yet the fan of the PSU is whirring away quite noisily. I wish Lenovo had paid more attention to the Advanced Dock 2503! The air that is coming out the back is cool and yet it sounds like it is at full-speed The dock had more potential (revision with some enhancements might have been good) however at the trajectory that Lenovo has been going on the dock is what it is- ingenious yet half-done right IMO. PCI x16 would be great! A beefier and less hypochondriacal power supply would also be welcome by me for sure! Also if anyone knows of a software or hardware solution to control the fan speed of the PSU turbine I am sure all who use the dock will be eternally grateful as well Still I was able to take the dock and put it to work with the kids at my parents place and they got to play some games with it- Bioshock Infinite, Tomb Raider (most recent version), Skyrim, and Resident Evil 5 for hours. Sheesh the kids must have sat there almost all day Saturday until I chased them off of the T500! I also got to work on assembling a T60 for my sis as an upgrade from an R52 she has had for 3 years. I installed Linux Mint on her R52 and is now a back-up lappy for her. I also wanted to put an Intel 6200 in my nephews X61s however it was not letting me install drivers for it. I will have to make a second attempt in August before my classes pick up again. I did want to test the dock with the T60 however I ran out of time and was back in uni town again after the 3-day visit at my parents. My reason is that as shown in one of my tests with the T500 with the CPU set at 'Low' the graphics still were running quite well. I figure that the T60 w/integrated graphics would run the same games just as well if docked. IOW the graphics card is very independent when it comes to constructing graphics and can do quite well even if the CPU is not the best. There are more variables to consider for sure! Hope the information has been helpful!

Future Considerations:

Test with aero off or on, Lenovo Power Management settings for Performance and Adaptive profiles, try 2gb of RAM versus 4gb RAM,test the X4500MHD and ATI HD 3650 GPU systems of the T500 (NOTE: will have to figure out how to streamline operation of switching from between all GPU somehow; 3 different GPU drivers does cause some issues).

[Medessec]

Anyone ever taken one apart?

Not far enough to get to the PSU... I'd want to do it, but my 2503 is too nice. I think I'll look into buying a separate, more cruddy 2503(if possible) and doing that.

Modding the PSU of the 2503 to accept higher power video cards has always been a bit of a thought in the back of my head... but the problem also lies with the space in the video card slot- cards that take more power are bound to be bigger, bigger than the card slot can hold. Of course, alongside modding the PSU of the 2503, you could also mod the video card slot to accommodate more powerful cards, but then it's not really a 2503 anymore is it...

The reason I think the Advanced Dock is getting more popular is that more lower-power cards are getting dirt cheap and bonkers powerful. We don't *particularly* need more powerful solutions and if we did, there's better ways of making it work.

I would like a way to mod the PSU, or just make it so it can handle moderately more power... but for now, I can't be moved enough to mess with my dock.

T500- W7 64-bit, Intel C2D P9500, Intel/ATI HD 3650 switchable graphics, 4gb DDR3 RAM, X-25m 160gb SSD (Main drive; OS), Samsung Spinpoint 500gb (Ultrabay Hard Drive; misc. files)

T500 using Radeon HD6570 through 2503 Dock

CPU: 1729.8, 2D: 461.7 3D: 352.3

These are some very curious results. Compared with my dock test results, so you guys don't have to flip pages:

T60/61 Frankie using Radeon 6450 through 2503 Dock
CPU: 1688.5, 2D: 144.0, 3D: 161.4
T60/61 Frankie using onboard NVS 140M
CPU: 1674.1, 2D: 260.1, 3D: 133.3

It seems your CPU is slightly more powerful than my X9000, although that might have something to do with my idiocy(I might have deactivated a core for cooling purposes, or the T61 cooler is struggling to let the X9000 roar, etc.) Your 2D results monstered both of mine, only finding itself comparable to my Clevo laptop and my roommate's X58 desktop... so the 6570 must have some better technology for handling that. Your 3D results were also comparatively better than mine, but still quite a bit below the W700ds' score, and of course to be expected... nowhere near the scores of my Clevo and the Desktop. Of course, it seems your card handled itself a lot better- did you get any PSU-related issues? Did the PSU fan whine when you were running the DX-11 tests? Or did you card run the tests smooth?

[emeraldgirl08]

The T500 is not too bad. It runs well enough and judging by the scores it is not ready for retirement just yet

To answer your question about my 2503 PSU fan. Yes it came on quite a bit. It seems even when I looked at the dock it would come on! lol Wish there was a mod for it!

I would like to run at least a couple tests of each card (ATI 3650, Intel X4500MHD, and the HD6570) and generate an average as I noticed an outlier or two. I was able to eke out one more set of scores from the Intel graphics and will get to ATI HD 3650 soon. The HD 6570 also needs a couple more runs as well (I only did 2 this morning). I am curious about the forum member who said he/she had a 2gb XFX card. I wonder what sort of scores that card might get?

[Medessec]

Yeah- I believe it was RMSMajestic, or one of his contacts discussing a 2503 with a XFX Radeon 7750 in it... the XFX 7750 as well as the AMD standard 7750, are the only cards as powerful as they are that are still single expansion slot and small enough to fit in the 2503. I'd imagine if ours suffer from power issues, those things have got to be stifled completely by the 2503's PSU.

I'm anticipating your score comparisons between the ATI 3650 and the X4500 Intel. The Intel scores should be particularly low, since if I recall... it's not capable of anti-aliasing 4x or higher, or DX11.

[emeraldgirl08]

The very last test with the tree islands and falling objects did not work for the Intel. There was a notification mentioning something about the anti-aliasing 4x indeed. I forgot to mention that. What I found strange about the testing was that the third test with full-screen airplanes flying around did poorly with the HD 6570 however the Intel graphics were faster by ~9fps I am curious how the HD 3650 will do. I think it will score close to the HD 6570. I mean the HD 6570 is really an entry-level card. I was not expecting too much. I did a benchmark test on Resident Evil 5 and it scored ~38 fps at medium settings. It looked quite nice on my monitor. Very sharp and fluid. I know when I ran that game on my HD 3650 a couple of months ago I had to place the settings lower than what the HD 6570 was using for the tests earlier today.

[QWERTY Andreas]

The reason i use the HD6850, is that graphics cards only run at a PCI-E 1.0 x1, which means that they suffer f
a perfomance loss of rougly 50%, compared to a fully fledged x16.

Its the same "issue" with the dock, resulting in a performance loss of these 50%.

The HD6850 would bottleneck my W500 at full speed, but since its only half speed, its about the perofrmance i wanted. I thougth about getting a 2503 instead - but the cards are so "slow" that it would not be worth it, taken the 50% performance loss into consideration.

[Medessec]

What I found strange about the testing was that the third test with full-screen airplanes flying around did poorly with the HD 6570 however the Intel graphics were faster by ~9fps

Did the Intel test run it without VS 2.0 and PS 2.0? If you run the test, it should post the notifications. If so, then that's why the Intel card is getting a higher framerate. The Intel card may not be capable of those visual technologies that the ATI card can do, but the test allows the Intel card to run the test without them. I believe the benchmark takes this into account when processing the final score.

It looked quite nice on my monitor. Very sharp and fluid. I know when I ran that game on my HD 3650 a couple of months ago I had to place the settings lower than what the HD 6570 was using for the tests earlier today.

It really is amazing what the newer cards can do in comparison to the older ones... especially the cards capable of DX11. When I upgraded my Clevo from the 285M GeForce to the 680M, it really was a case of "why do I need anything else." While the 285M constantly struggled with any sort of rigorous graphics-intensive activity, the 680M just walks over everything like it was nothing. And being able to play all of my games on High to Ultra settings, and in DirectX 11, is a luxury I could never ask for with the 285M.

[emeraldgirl08]

The reason i use the HD6850, is that graphics cards only run at a PCI-E 1.0 x1, which means that they suffer f
a perfomance loss of rougly 50%, compared to a fully fledged x16.

Its the same "issue" with the dock, resulting in a performance loss of these 50%.

The HD6850 would bottleneck my W500 at full speed, but since its only half speed, its about the perofrmance i wanted. I thougth about getting a 2503 instead - but the cards are so "slow" that it would not be worth it, taken the 50% performance loss into consideration.

This is what I have read about but wanted to see how it translated into my particular setup. Have you got any numbers for the performance of your HD 6850 in your setup? Just curious. If yes how are those numbers in comparison to other benchmarks with the same GPU and other similar performing hardware?

Did the Intel test run it without VS 2.0 and PS 2.0? If you run the test, it should post the notifications. If so, then that's why the Intel card is getting a higher framerate. The Intel card may not be capable of those visual technologies that the ATI card can do, but the test allows the Intel card to run the test without them. I believe the benchmark takes this into account when processing the final score.

Not sure what the difference between VS 2.0 and the PS 2.0 is Could you elaborate? As far as notifications the Intel test did produce a notification and that was for the very last test w/the tree and the camera panning around it while objects w/smoke were falling from the sky. That did not run on the Intel integrated graphics. How I ran my tests was with Performance Test 7.0 was using the menu>> Tests>> Choosing either 'CPU,' '2D Graphics,' or '3D Graphics' >> All (which meant the selected test would run the hardware through a sequence of specialized tests; e.g., CPU would run through integer math, floating point math, find prime numbers, and 2D would run solid vectors, transparent vectors, complex vectors, etc.). I am curious however doubtful this would change anything but I would like to download the driver for the HD 6570 directly from AMD. I am using the latest driver from the MSI website for my HD 6570. I will add that to my future considerations as well. It is premature for me to make a conclusion but if jumping to one I would say there is some gain in performance. After all the more recent hardware and software for the HD6570 might give it some type of edge in performance. I also get the benefit of HDMI w/audio as well. At the end of the day this is a project so for anyone wondering why I am investing time, my laundry money, and energy into this it is to mainly to satisfy a curiosity and to get some actual numbers as well

[QWERTY Andreas]

This is what I have read about but wanted to see how it translated into my particular setup. Have you got any numbers for the performance of your HD 6850 in your setup? Just curious. If yes how are those numbers in comparison to other benchmarks with the same GPU and other similar performing hardware?

The HD6850 will get 20000 in 3Dmark06 when running full speed, in a desktop.
I get 10000 with the eGPU setup.

The FireGL V5700 get around 4000 IIRC.

The thing about the external PSU is no big deal for me. But if somebody wants a nicer solution, connection wise, they can just use a PCIe raiser with the dock

[Shredder11]

If anyone has any help or advice with a HD 6570 1GB card in the advanced dock, have a look here at my post:

http://forum.thinkpads.com/viewtopic.ph ... 34#p749134

[86turbodsl]

I've been thinking about this recently, as i have several 2503's sitting at home. The constraints on video cards in this dock seem to be centered around
power usage and size. That got me to thinking about a PCI-e extension harness and vidock type setup, but used from the dock.

What are the bandwidth constraints running a remoted card from the dock vs running an expresscard54 expander?

I'm assuming pcie lanes in the laptop expansion connector will only be one lane, but i don't know that for sure.

Any wisdom out there?

[RMSMajestic]

I've got a low profile HD7750 in my 2503 dock.... it feels better than the other GTX645 that I've tried
http://forum.51nb.com/viewthread.php?tid=1478474
The GTX645 I used was a physx acceleration card on my desktop together with a HD5870. Although it's a single slot card, it can't fit in 2503 ultrabase
I had to modified the 2503 ultrabase a bit for better airflow for the HD7750. Otherwise it will overheat.
On my desktop, then HD7750 reaches 80 degrees celcius when full loaded. In 2503ultrabase, temperature was 5 degrees higher

[coolcat37]

I wanted to put a 7750 in mine but the prices are too high for this. 180 euro seems to be a bit steep.

Has anyone experience with the advanced dock in combination with a video editing program (specifically Adobe Premiere in Windows 7) ? Does the external video card speed up the render process ?

and anyone experience with a graphics card in the advanced dock whilst using Linux (Mint)?

[micrex22]

Alright, I've got some actual testing done. And I think I've determined that the underperforming qualities of my Radeon 6450 are stemming not from the card itself, but from the 2503, probably it's PSU.

The 2503 is actually PCI-e x8 or something and not actually x16 (take a look at the traces inside the 2503 on the slot), so PSU aside, the card will never run at the full x16 bandwidth.