T61p's Nvidia 570M, the 14" card is not the same as 15"

Besides the obvious fact that the T61p 15" has 256 MB dedicated video memory over the 14" one with 128 MB, this site:

http://forum.thinkpads.com/viewtopic.php?t=48287

indicates that the 14" T61p has a underpowered Quadro 570M with only a 64 bit memory bandwidth. Quite disappointing.

It makes me wonder why Lenovo would do this because they are charging more for a 14" t61p even though the video card is equivalent to a 14" t61 with Quadro 140M.

 

Well, a 8600M-GT with 64-bit interface would still be more powerful than the 140, because of the number of stream processors. But if this is true than it does seriously suck.

 

D A N G I T ! ! ! !

Why, oh why, lenovo?

 

Well I am canceling my order right now. Thats is sad by Lenovo. I knew something was wrong with this picture. Grrr

 

this is true my 3dmark and vista scores are very similar to the quadro 140 card. Lenovo keeps propagating more and more scams. what a crappy company

 

How is this a scam by Lenovo? It's nVidia's chip...

 

I dont even see the 256mb being sold anymore. And didnt the 3dmark06 hit into 3k? In the review? Which is no slouch.

 

t1mmer - because they don't specify that the card isn't running at full spec - It's not exactly a SCAM, but it's pretty cheap.

 

So does this mean the 14" is not going to be able to handle a lot of the future games?

How will it par with other the 14" notebooks on the market..?

 

It's got plenty of pixel-pushing power, but it's going to be hobbled when dealing with a things like high resolutions, anti aliasing, and large textures. I'd bet you'd have to turn off AA and crank the textures down to Medium on a lot of games with the lower memory bandwidth.

 

shouldn't be buying a quadro notebook to play games anyway.

 

Thats for the 15.4" with the 256mb, not the 14.1" with the 128.

 

I see people throwing this around a lot and it just doesn't make sense to me. It's the exact same chip, can use the exact same drivers with a bit of effort, and a lot of us have ordered 540M notebooks for the same or less than an equivalant 8600 GT notebook would cost. Hell, I've done just the opposite on my desktop. I soft-modded a 6600GT to look like a Quadro in the drivers and enable graphics acceleration in CAD. Again, the cards are the same across the consumer and professional lines, so as far as I'm concerned it shouldn't really matter.

 

sorry dude it has been confirmed. Look around. And the fact Lenovo makes no mention of this anywhere really pisses me off. Just like the Dell 8600GT problem. Same thing. Its shady and not right.

 

No, this is not even close to the Dell problem. Dell was selling a card that was fully up to nVidia's specs (as nVidia's documentation gave specs for both the DDR2 and GDDR3 versions of the card). However nVidia makes no mention anywhere of their being a 64-bit FX 570m. This is similar to those who got screwed with the 128mb 64-bit x1600. However, I'd still like to see gaming benches of this. If it's up to par with the 8600M-GS, I may still get it.

 

I was speaking to the generality that Quadros aren't good for gaming, not that this particular Quadro was gimped. To all indications the 15.4" t61p has what amounts to a rebadged 8600M GT. The 14.1" t61p version is pretty crummy, I'll agree wholeheartedly.

 

That's not really true though. In general, quadro cards are better than geforce cards at everything, as they have higher quality hardware. With the same drivers performance is at least equal.

 

Actually it used to be true. Older desktop Quadros (I think up to the 6-series derivatives) used slower RAM, since raw memory bandwidth isn't as important in technical work. They still charged a 200%+ premium for them on account of the more thorough binning process (like Xeons vs. Pentiums -- only real difference is the quality control). Quadros still perform more poorly in games with Quadro drivers, because they're not optimized for gaming performance. Still, it's not that hard to get regular drivers working on them, and after that there's not really much of a difference.

 

Yes, that's correct. I think the change started with the 6800 and it's quadro equivalent? Though I could be wrong about that.

 

Okay, I just reread the rivatuner's output from the thread linked in the first post. At stock clocks (with quadro drivers) the 14" and 15.4" T61p should perform the same. The 14" has half the bus width (in bits), but double the memory clock (in mhz), so the memory fill rate (in bits/sec) will be the same. However it's true that this limits the overclockability of the 14".

And someone finally posted 3dmark06 scores, it appears that the 14" is slightly behind the 128mb MBP (which has the card we thought it would be based off of), but not by much. I would seriously be interested to see gaming benches of this lappy, at 1024x768, 1280x960, and 1400x1050.

 

Well mine is coming in a week or so and when i get it i will be sure to install the Geforce drivers and run any tests you guys want.

 

when running ArcMap on my T60p with the 5250 compared to the T61p with the 128mb 570 there is a huge difference. The textures and layers on my T60p are much quicker to refresh and manipulate, the T61p is sluggish and takes a while to refresh. I am really dissapointed with the the T61p, this card is definitely not even remotely close to the 15.4 570.

 

You're right, it's not. It's performance is still best in class for that size though.

 

Bottom line: Is it still better than Zepto 6224W's 8600GT GDDR2 when both are at max TurboCache?

Hence, is it still the best GPU for 14" so far?

 

Is one running XP and one running vista?

 

both are running xp the T61p is running XP on the 80gig 5400RPM drive that came with it, but I also have a 7200RPM 100gig hard disk that I use to run vista on it as well. I cannot get Arcmap to work on vista so I have been using this method to get it to work. My T60p is running XP.


Quick question. I got 3dMark06 to run but I cannot view the score everytime it finishes, I click on view your results online and it says submission failed any ideas on how I can get my score to show?

 

that probably why.. your forcing a dx10 card to run on a dx9 platform

 

The results posted in the link in the original post are now said to be incorrect.

 

It's equal to it in gaming performance. The GDDR3 64-bit should essentially perform the same as the DDR2 128-bit (assuming the same drivers and nVidia clocks). However I'd still choose the T61p due to the better portability.

 

That guy was just stating the nVidia spec'd clocks. No one has actually posted ntune results yet (to the best of my knowledge).

 

How so? Zepto's 6224w is a 14"

Also, Asus F8Sv is a 14" with 8600GT

If it comes down to looks, T61p will really fall flat.

What about the 14" 570m VS 8600GT GDDR2 when both are safely overclocked as much as possible?

 

I haven't seen overclocking results of the 570m yet, so I can't really say.

The T61p is lighter with better battery life.

 

These results are really going to decide for me whether or not I'll be getting a T61p.

Yea, looks aside, I'm really interested in using the Ultrabay Battery to get maximum battery life in all comparable 14" laptops.

Although, I heard the UB dies quite quickly because it gets drained first before the 6/9 cell is used, correct?

 

Someone really ought to take the time and install the GeForce drivers on the thing and run the appropriate benches..

 

No kidding...

 

If what you said is true I feel less disappointed and may still want to buy this notebook because it's so difficult to find a 14" notebook with a decent graphics card (there are few choices), but yeah I'll like to see some benchmarks.

 

So if I'm understanding this correctly, the 128 is clocked double the speed of 256 in order to perform equally well (throw the memory factor into the water for now), meaning the 256 should be able to safely clock doubled, and then some more, whereas the 128 can only clock that "some more" before maxing out.

 

well...that's a bummer....64-bit interface >< i doubt it'll be able to run crysis well now

 

Wiki shows both cards have a 128-bit memory interface.
Wiki Article

It also shows the 140M as a 128-bit, while nVidia shows the 140M as 64-bit and the 570M as 128-bit
140M
570M

Comments?

 

I'd go with nVidia stats over wiki stats any day of the week.

And also, I think that the general conclusion is that benchmark-wise, this 570m scores around the same as the 140, which really blows.

In terms of gaming then, I think it's ranked:

1) 8600GT GDDR2
2) FX 570m
3) NVS 140m

This is in terms of 14" GPUs, of course.

 

It's odd that NVIDIA wouldn't mention something like this. Could it be that some of our information is not correct..?

 

2088 3dmark06 I am overclocking it now to see what I can get

 

2088 is the 14in with the 128mb nvidia570? What cpu? The review had the 256mb version on the 15in with the 2.4Ghz processor and it broke the 3k mark.

 

T7300/ 14.1/ T61p with the Nvidia FX 570m/ 4gig ram/ Vista Business 32

Overclocked at Core Bus 611/ Memory Bus 810 I am able to get 2585

 

No, the 140 is just a downsized 8400M-GT with 64-bit bus width, I can't imagine it would ever break the 2000 mark at stock clocks, so this card is significantly better (considering that 3dmark06 likes 256mb of vram). The differences between the nvs140m and this FX570m is that the latter has double the stream processors, with higher stock clocks and greater overclockability.

That's actually not too bad, considering that 3dmark06 is heavily biased towards having at least 256mb of VRAM. I'd be seriously interested to see some gaming benches with Geforce drivers. Right now I'm guessing that at lower resolutions this 14" 570m will be the approximate equivalent of the DDR2 8600M-GT.

 

What game benches would you like to see, I will try them when it arrives..?

 

What games do you have, or are planning to get? Things like a CS:S stress test, a FEAR benchmark or an Oblivion test or even a Bioshock test would be useful. Really any modern game would work.

 

Well after half a dozen 3Dmarks the T61p 570 14" pulls a consistent ~5400 3dmark05 score, with regards to 3dmark06 it pulls ~ 2300 consistently.

Basically it outperforms the reviewed inspiron 1720 with a 8600mGT w/T7500 (I am running a T7300) in the 3DMARK05. With the 3DMark06 the tables are turned and the 8600M/GT in the 1720 mops up the T61P's 128mb 570.

 

like odin243 said, 3dmark06 favors 256mb cards, so that could be the reason. do you find the card disappointing in real life applications?

 

That's quite expected, because 3dmark06 is highly optimized for 256mb cards. The higher 3dmark05 score is encouraging, since that program tests for 128mb cards. That seems to imply that on lower resolutions (as we knew the 128mb couldn't handle real high resolutions) the 570m is still the better choice than the DDR2 8600M-GT. Also, are these scores with all stock settings and drivers?

And, do you have any modern games that you could test with?