NVIDIA Quadro FX 3500M VS. NVIDIA Quadro FX 1600M in mobile workstations

Hi all


I'm just about to purchase a mobile workstation with the NVIDIA Quadro FX 1600M but now I found out that there are other workstations using NVIDIA Quadro FX 3500M which seems to be better, so now is causing me some doubt.

So I found this little chart . Seems like 3500 is outperforming all the rest. However, I read elsewhere that Quadro 3500 and 2500 are older generation models, in other words, 1600 is newer generation and will deliver pretty much the same, if not better, performance. I think it has to do with 1600 having DirectX 10 functionality, but I do not intend to use Vista (I want XP x64), so, should I care about this?

In other words, can anyone help me out with interpreting those charts? I specifically have no idea as to what "PowerMizer Version" means, "C Programming", "Graphic Memory Bandwidth" and "Memory Interface" ???

Then let me know which card you think it's better. It will be performing with 4GB of RAM and 2.6 Duo Intel processor running on XP x64 Pro

Please?


Thanks, lots

 

Umm, did you see this http://www.nvidia.com/object/IO_11761.html This may answer all your questions. Oh by the way, are you using the 6600 desktop processor in your workstation or the intel extreme for laptops?
Scroll down and you'll see that the Quadro fx 3500m is the best card, however it doesnt have dx10 capablilities while the Fx 1600 does.

 

power mizer i suspect is power control feature saves battery life.
C programming no clue
Graphic Memory Bandwidth: the more you have the more detailed images and faster display of those images
Memory Interface: just what it says, 256 bit bus, 256 twice as much theoretical bandwidth as 128
Since you wont be using DX10 3500 is going to be better.

 

ouch... that's not what I wanted to hear...

if I'm getting this right - the only way to experience DX10 is if you are using vista, right??? damn, this 3500 is so much more expensive... is it also that much better?

 

As near as I can figure the 3500 is comparable to a 7950GTX, and 1600 is comparable to 8700m GT, Yea the 3500 is better but 1600 is very good also and may be enough for what you need, if so you won't see a difference it is the best DX10 out there for notebooks so it has to be pretty good.

 

would quadro 3500 outperform 7950 GTX?

 

(for animation/graphics/rendering tasks, not game play, sorry)

 

I think 7950GTX is optimized for game play (DX) 3500 is optimized work (Open GL). So overall for what you need the 3500 was designed for that purpose and therefor better suited. That is why work stations have them and not 7950GTX. The drivers are what make most the difference. And I do not know for sure it the 3500 is actually based on the 7950 but I suspect.

 

powerpack is correct on that last point.

To answer Uxion's question, the 6600 is 2.4 GHz on the desktop side, and the next step after that is 2.66 GHz (6700). So if it's exactly 2.6 GHz, it must be the Core2Extreme X7800, unless milenica got an early order of the T7800. Either way, that is one very nice, powerful machine.

 

Apollo - I am about to order an HP machine with the T7800, but the Quadro FX 1600 card (running on XP though)... Taking the Quadro FX 3500 would cost me additional.. well, say $900, which I'd have an incredibly hard time pulling off... Also, that configuration would come only with ntel® Core™ 2 Duo Processor T7600 (2.33GHz/667MHz/4MB) Do you think it's still a good thing to get the first option, with FX 1600? I understand it has the DX10 option, but unless I'm using Vista (Which I won't) - I won't get much use from that. Just trying to figure out how much I'm losing/gaining by opting for a better processor, but less powerful GPU.

thanks

 

Your saving $900 get the 1600 unless you know you need 3500. What do you plan on using this work station for because both are beasts and would exceed the power needed by 99.9% of ppl (non gamers). An old saying "if you have to ask you don't need it"

 

I would be using it for heavy 3d modeling, rendering, animation and designing presentations + portfolio books using Adobe suite. All other CAD work too

 

What was your previous GPU? And how did it perform with these tasks?

 

Errrhm I have right now NVIDIA GeForce FX Go5600, running on 1GB RAM and Intel Pentium 4CPU 3.06GHz... This card was not so bad, I suppose, though this configuration is 4 yrs old. Right now I do have problems handling bigger files in Rhino (3d modelling program), even though I have pretty good skills in reducing file sizes and keeping unnecessary stuff out. Rendering times were not stellar at all (had to use my bf's computer a few times, because mine would not pull it off)... And it's just old, you know, and I want something new and fast that I won't feel guilty spending over 4k bucks on. But, I also want to really use it, rather than buying something that's great, but the system is not fully utilizing it...

grrrrr

 

The 3500m is going to be at least 25% faster than 1600m in any CAD work, possibly going up to 50% faster or more at high resolutions. Which one would be right for you depends on how hardcore your needs are.

Edit: I would say your best bet would be the 1600m, it should be fine for you. If you need more mobility, even the 570m should be fine for your needs.

 

I agree with Odin's edit, but only you know your needs. Both systems are a major step up from what you have, I think you would be happy with either.

 

thanks a lot for all opinions.

I think it might be a safer choice to go for a better processor and the FX 1600, then later upgrade to a better GPU.
However, I remember reading somewhere that you CAN'T upgrade your graphics card on a laptop!?!?! warranty issues too...

can anyone confirm this?

 

ah, well, so it is impossible...

link

 

Most of the professional applications are CPU limited not GPU in contrast to games. In other words, the CPU is the primary determining factor of the rendering speed. And, the difference between the high-end Quadros would be significant only in some high-end Visualization applications (talking about miltiple displays, high-resolutions, complicated models and real time rendering) otherwise, you wouldn't see much of difference and 900$ for the Quadro FX 3500M would be rather waste of money...

The HP 8710w has replaceable video card but you can upgrade it only if HP offer another video card option for this notebook in the future...

 

heavy architectural models/CAD and rendering don't fall into that category? what about previous advice of using 3500 rather than 1600 on the cad applications...? totally confused.

thanks for the mentioning the 8710w, I am still leaning towards it.

 

The "any CAD work" part really makes this post just as ignorant as it can be...

And now again, the CPU is the bottleneck for most CAD applications.. that's it at least for now. Otherwise, there are high-end visualization and other applications that can make use of top-of the-line video card... but that's a different story.

It's nonsence considering that there are CAD applications that don't even require a video card since are mostly 2D oriented, and other that just have some 3D features...

HP use some proprietary (it's not confirmed but it's not a standard MXM most probably) format... Either way, the card is replaceable unlike most notebooks which have GPUs soldered to the motherboard. However, I don't know whether HP will offer another video card or not, they have done that before but now I don't know...

Here are some workstation benchmarks...

Notebooks/ Applications	HP 8710w, Quadro FX 1600M (512MB), based on the 8700M GT Core 2 Duo 2.4Ghz, 4GB	Dell Precision M90, Quadro FX 2500M (512MB), based on the 7900 GTX Core 2 Duo 2.33Ghz, 2GB	Dell Precision M90, Quadro FX 3500M (512MB), based on the 7950 GTX Core 2 Duo 2.33Ghz, 2GB
3ds max	32.61	30.35	30.02
CATIA	38.64	37.1	36.8
EnSight	30.8	25.7	25.51
Lightscape	33.15	31.35	31.04
Maya	59.52	136.3	133.7
Pro/ENGINEER	35.55	36.81	37.07
SolidWorks	45.07	52.73	52.91
UGS Teamcenter Visualization Mockup	17.33	13.52	13.5
UGS NX	16.8	16.97	16.9

...so I said what I said.

The HP 8710w uses the new Santa Rosa chipset which has slightly positive effect on performance in some workstation applications, so that's way some results are slightly higher...

 

The card is still faster and more powerful Dreamer, whether that speed and power is utilized or needed is a different matter.

You should really include CineBench OpenGL results, as the one's you posted (except for Maya) appear to be mainly CPU benches, not workstation GPU benches.

 

That's quite a stupid post, so this card would be faster for playing Solitaire as well, it doesn't matter that the speed/power will not be utilized at all and you would pay 900$ more for it... you're just pathetic.

Btw, faster for what? CAD applications? Which exactly?

Not before you show a direct connection between CineBench OpenGL results and performance in certain workstation applications.

The bechmark I posted is called SPECviewperf, which is the most polular workstation GPU benchmark in case you haven't heard.

http://www.spec.org/gwpg/gpc.static/vp9info.html

Keep sharing your ignorance with us, I'm really having fun...

 

Yes, this card would still be faster if you were playing solitaire as well. I didn't say "faster for what", I just said it's a faster and more powerful card, which is true. And I never said the OP should pay more for it, in fact if you'll bother to read my posts I said:

I'm not sure what you're asking for? It directly tests the suitability of a card for handling Cinema4D applications, and the Open GL show how well the card can handle enhanced OpenGL displyas in the Cinema4d viewport. Any modern workstation application uses the same type of OpenGL display for viewing 3d work.

 

I will..

So you said "faster for what", no?

And I said that's nonsence, which it true. Savvy? Or you want to tell me about CAD applications in which the 3500M will be really that much faster?

Tip: CAD applications/work being the keyword.

I'm asking you how this is connected with the CAD applications?

Wait, I got it... you haven't a clue what a CAD application is, right?

*sigh* I'm not surprised anyway.

 

guys, I do appreciate all the information, I just would sincerely hope that all these posts are done in nothing but good (competitive!) spirit

Alas, I have made my choice and placed and order for HP8710w with Core 2 Duo 2.6Ghz, 4GB. Thanks a lot for posting the benchmarks, although I have no idea what the numbers mean (do pardon my unfathomable ignorance, please I will only assume larger numbers mean better performance, in which case I'm surprized to see the results for Maya... ehm, why so low? In any case, wish me luck with my new workstation.

Many thanks, again, to all of you!

 

No, it's not nonsense. The real-time display and manipulation of the viewport for any heavy 3d modeling or rendering with CAD programs will be heavily GPU dependent.

Because CAD applications use that type of OpenGL display, as is shown directly in things like ArchiCAD.

Congratulations on your purchase, that is a very powerful machine and I'm sure you will be extremely happy with it.

 

In short, when you're going to claim something at least have a way to back it up, otherwise, you're just looking like a fool. Plus, you haven't a clue what you're talking about.

True.

Maya isn't a CAD application...but don't worry about that.

Good Luck.

 

Actually, everything I've said so far has been accurate. What exactly haven't I a clue of? Do you disagree that: "The real-time display and manipulation of the viewport for any heavy 3d modeling or rendering with CAD programs will be heavily GPU dependent."

 

This is a general statement that means nothing.

Lol...

Well, I've been ignoring your posts for months now and I think I will keep doing the same... since it's for a reason.

This statement is nonsense, so either provide evidence to prove it in your next post or you are an idiot.

 

AKA you don't want to admit that it's 100% accurate.

If you don't care to back up your claims, and instead make general insults, then I think we can see who's posts are worthwhile.

How is it nonsense? The 3500m is physically faster than the 1600m, in any application at any time. It executes more instructions per second. How is this too hard to understand? Whether that speed is needed or useful is, as I said before, debatable and depending on circumstance.

 

Nope, it's just a post that every 13 year old member of this forum would make when he has nothing to say...

Haha, I backed up everything what I said and provided benchmarks for the most popular workstation applications, which shows that I'm right. You're the one who didn't provide any evidence to back up your words... and if you don't do it, then yes you're an idiot.

How is that hard to understand that certain applications cannot use that much of power or at all, and it's not debatable but it's simply proven fact?

Still waiting for Solitaire benchmarks....

In the end, there isn't much of debate in here, so either try to back up you post with benchmarks (remember: CAD applications) or you know...(see above)

Now, I would like to see benchmarks showing the performance of the 3500M and 1600M in all of the CAD applications since you said "any CAD work" and you seem to be sure in what you're saying... so prove it...

And, I don't need general statements or numbers, I want to see benchmarks like those I posted, which are different viewsets that simulate real-use case scenarios for the applications that are tested... isn't that simple?

 

Is it just me or is it really uncommon to see fighting between two people, especially with tons of Rep each?

 

First of all, not many 13 year olds that I know are into CAD, but disregarding that, it does not change the fact that my statement was a 100% factual representation of when a high end GPU would be needed in CAD work.

I have never once said that the applications used by the OP required a 3500m, nor have I ever denied that certain applications don't require high end graphics. However it's also impossible to deny that cards like the 3500m are developed for a reason, that reason being that some CAD work does indeed require high end graphics cards.

That's fine, and I've been compiling some data over the past several hours, here you go. It's the SPECapc for SolidWorks 2007, which was the most relevant standardized benchmark I could think of that test real world situations.

 

Kind of a battle of the Titans, I would only like to say a Corvette doing 60 MPH and a Civic doing 60 MPH are both doing 60MPH and will arrive at the destination at the same time (assuming they left from the same place at the same time). I will now go away while the big boys fight.

 

I haven't seen facts yet... just words..

Sure, the 3500M is here for a reason (even more than one) but this reason isn't CAD in most cases.... There are DCC, NLE, VA, SA etc. applications that's the reason.

You have, you said that the FX 3500M will be faster in any CAD work, which basically includes all CAD applications, which isn't true, not only because most of them are CPU limited but also because there are 2D CAD applications, which makes your statement ridiculous....

OK, but it tests only one application, plus you can see a difference in my table as well, so what exactly are you showing in this way?... nothing new anyway... Do I need to wait another few hours for your next discovery?

I don't know but well, try to come up with something better next time, again, you're claiming that the FX 3500M is at least 25% faster in any CAD work... prove it... (btw, my table doesn't think so... lol) and I will ask you to prove these 50% or more at high resolutions as well.

 

The FX3500M is faster than the FX1600M any time, in any application, including all CAD applications. You seem to be confused between the card being faster, and that speed translating into the application running faster. I never claimed, nor would I ever claim, that a faster graphics card will necessarily make your CAD application perform better, but that doesn't change the fact that the graphics card is still physically faster.

I'm merely providing support for my statements with a performance analysis from a real-life CAD application. Your table does not truly show the graphics difference between the two cards, as you only have general scores listed, not the OpenGL graphics subscores from the individual SPEC tests. And in case you didn't realize it, my benchmark shows the FX3500M being 27% faster in the OpenGL display portion of the test, and it shows the FX3500M system being 9% faster in the entire SolidWorks application, despite the 8710w outperforming it CPU-wise and File transfer-wise. I will consider this validation of my claims unless you have any other specific tests you want me to run to show the difference.

P.S.: I will try to show the difference at higher resolutions, though it's a little difficult with standardized benchmarks.

 

That's well written nonsence, I must admit...

Let me see, the FX3500M is faster than the FX1600M any time, in any application but a huge amount of applications don't run faster on this card... how can it be faster in any applications then...

Sorry but this is just idiotic.

Plus, who asked you whether the card is faster in general? The OP asked which card would be a better choice, then you made your nonsence post about the FX3500M being faster in any CAD work and confuced him....yep, that's idiotic.

And, I will never read similar idiotic posts again I swear... that was. I'm not here to deal with people's stupidity at end.

Haha, if you don't like my table and especially weighted geometric means, than post a "better" table and prove that this table "does not truly show the graphics difference", if you don't do it, then that means that you accept these results since you can't prove the opposite.

I realize that you posted results from a specially chosen application... that's about, and that certainly isn't representative for all CAD applications.

Well, you may consider this for whatever makes you happy, but I see it like nothing...

Oh, and you can consider the Earth being flat as well, I don't mind that.

And then again, I'm not gonna read any of your idiotic posts any more, so unless post a "better" table...I'm done with this thread.

 

There is a difference between the statement: "The 3500m is faster than the 1600m in any application" and the statment: "Any application will run faster on the 3500m than the 1600m". I said the former, but I never said the latter.

Perhaps you would care to actually read my suggestion to the OP?

If you take the time, you'll see I very clearly said that the extra speed of the 3500m would not necessarily be beneficial for his uses.

Well, I can prove it with your own words, as those tables are indeed "weighted geometric means". I.E. they are taking into consideration many factors besides graphics differential between the three systems, thus those scores are not a good representation of said graphical differential. Also as an aside, if you believe the bolded statement then you seriously need to brush up on your logical arguments. An unproven negative does not imply a positive.

Good riddance.

 

:laugh: :laugh: :laugh:

Nice try, I guess, when people feel hopeless they are trying to hide themself behind semantics...

Sadly, I don't see the difference, and the OP didn't see it as well...

Yeah I "see" that, and the OP saw it as well.

No comment.

No, you cannnot prove anything, and we just see another general nonsence comment...

Now, it's simple, can you provide results for the same applications that would be "a good representation of said graphical differential" according to you?

Otherwise, you just're just saying that something isn't good but you don't have anything better to offer... how do you know that it isn't good then?

Plus, you don't expect that people will take your word as gospel, do you?

Also, I don't care how you feel about the benchmarks I posted, I will let people read them and decide for themself, and you can keep your feelings for yourself and your diary of course.

I believe that if you don't care to back up your claims, then your posts are just rubbish that clutters the forum and I don't feel need to read them, and if your next post looks like a pathetic essay and you don't provide any evidences again, then I'm not going to respond anymore and you can take that as you wish, as I said before you can consider the Earth being flat as well I don't mind that... hopefully, you will find someone who will take you seriously.

 

Sorry to the OP but this has degenerated into a flame war. Please try to consider that these threads are designed to help people choose which notebook they'd like. If you wish to discuss these issues further please do so in a dedicated thread.

 

Keep these ridiculous arguments off the forums for future reference. You know a thread is going to get closed if there is constant back-and-forth bickering. I don't necessarily mind an argument as long as it's constructive and not dealing with personal insults, but this thread was far from that. We have no tolerance for personal comments.