8600M GT or 4670 ati?

Which is the better graphics card the 8600m gt on the m1530 or the new ati 4670 on the xps 16?

 

The 4670, by far. The 8600M GT GDDR3 on the XPS 1530 is slightly better than a Mobility Radeon 3650, but the 4650 is much better than the 3650, and the 4670 is a bit better than that.

It's essentially comparing a upper midrange 2007 card to a upper midrange 2008/2009 card, and there definitely was considerable progress, especially by ATI.

 

I'm a little pissed I really wanted the xps 16 but decided to go with the m1530 from the outlet because of the heat issues. I wish I would have waited but I ordered the laptop 1 month ago which means I cant return it for the xps 16 correct?

 

So instead of ordering a laptop that gets a little toasty, you ordered one that has a faulty and destined-to-fail-prematurely GPU?

 

^^^^what he said..

you went from a laptop with a beautiful screen to a butt ugly one. wow bad choice.

 

what ???

i didnt understand this thread.... so u bought the laptop...and now want to know what is better

 

+1 for the 4670. You should have searched around the forums for the 4670 update.

 

I wouldn't say that the m1530 screen is but ugly. It's actually a very nice computer and very happy so far. I wasn't aware of the problems with the nvidia card till I purchased it but so far so good. I didn't know about the 4670 card so I pulled the trigger with the m1530, great price with 15% outlet discount.

 

HD4670 is much better than 8600M GT. HD4670 has also GDDR3 memory, like the 8600M GT or HD3670.
HD3670 (old version of StudioXPS 16) is about the same as powerful as 8600M GT in the XPS m1530.

 

3670 is far better then the 8600m gt.

Actually, I don't think the 4670 is that much better. I think it performs so much better is because it got the new drivers (so does the 3670), while when the 3670 was reviewed it used crap drivers.

 

8600m GT < ATI 3670=9600m GT< ATI 4670

 

in avg scores

8600M GT scores:
3dmark05: 6175
3dmark06: 3347

HD4670 scores:
3dmark05: 12614
3dmark06: 7308

so it seems, the HD4670 has more than twice performance than the 8600M GT has, no wonder, 8600M GT is a pretty old card, 4670 is kind of new.

 

Haha..thats so far from reality. There's a huge hardware difference between the two.

 

This is mistaken.

THe 3670 is about the same as the 8600M GT (GDDR3) in the M1530. There is no massive difference between the two, in fact I would not be surprised if the 8600M performs better in plenty of games.

The 4670 is a lot better than the 3670, I don't understand why you would say it is 'not that much better'. The specifications and performance difference is significant.

 

yep im with VC on this one, the 8600gt is a powerful card, its almost identical in performance to the 3670, but the 4670 beats it because its almost 3 years of difference.

and for those who said the m1530 is ugly; go see IBM laptops and then come and tell me the m1530 is ugly.

the m1530 is a very old model, but still tackles most of the games out there with ease, an awesome buy

 

Nah. A Mobility 3670 performs significantly better then a 8600M GT (no, actually it beats it pretty hard down in the ground). I already have posted a youtube video with some guys playing Crysis Warhead on gamer and probably 1024x768 resolution, and enjoying roughly 20+ fps (and the GPU used drivers from April/May). An 8600M GT can only dream of such.

As for my second point: The 4670 isn't actually that much of an increase specification wise (only more sharers). However, it has new drivers installed. Back in 2008 the ATI 3670 had very bad drivers, so it performed horribly. And you people are comparing its performance from 2008 to the 4670's performance today.

NOTE: I did not say the 4670 is worse or not any stronger then the 3670. I simply said it isn't that much better then everyone thinks .

 

I can't but help feel you might be a little biased to say the the M1530 is awesome, as sometimes it's hard to admit your stuff isn't awesome anymore.
Looks are purely preference. I'd rather have a Thinkpad any day of the week over an XPS 1530.
The 8600M GT was midrange in its time, and it's already fairly dated. It will still handle new games on medium settings which is cool, but let's face it- the 8600M GT is a faulty chip - so it's hard to legitimately consider it an "awesome buy" or even a "reasonably good buy" with that knowledge at hand. There are much better laptops for the money, the M1530 is an outdated laptop.

 

The 4670 effectively doubles the performance in most games coming from a 3670, this is a fact regardless of which drivers you use. Going from 378 million transistors and 120 unified shaders to 514 million transistors and 320 unified shaders is a huge jump. Not to mention the 3xxx series is an older piece of tech compared to the 4xxx series. So yes, it IS that much better just as everyone thinks. I do agree that the 3670 is a noticeably better card than the 8600M GT, it gives up to 15% better performance in many situations.

 

And ATI used different technologies when making them. Anyway, based on benchmarks (the only good way to measure performance) the 3670 didn't perform even remotely close to half as bad as the 4670. You are claiming that the 4670 can manage 40+ FPS on Crysis Warhead Gamer settings and 1024x768 resolution and 80-120 FPS on Wow maximized except shadows at 1920x1080 resolution. Which it can't.

I know you want to defend your buy, but bad drivers have allot to do with the previously poor performance of the 3670.

 

I don't need to justify my purchase at all, especially for a middle of the road gaming card (if I wanted a true gaming laptop I could've gotten one for less money). I'm stating the facts and those are that the 3670 in most cases offers only 50% of the performance of the 4670. Driver revisions don't close a huge gap like that unless there was something significantly wrong and that isn't the case. You're trying to argue that the notebookcheck numbers are off simply because of a driver revision difference and that's laughable. The 4670 is a middle of the road video card and the 3670 these days is bottom end. I'll be happy to post up apples-to-apples benchmarks once the first set of 4670 SXPS 16's arrive. I'm sure you will have more excuses even then.

It sucks you can't get the 4670 where you live else I'm sure you'd have jumped on it. Rationalizing a purchase when there's something clearly better available is a pretty crap position to be in, I feel for ya. BTW I know EXACTLY how the 3670 performs, I had a Studio XPS 16 previously and I pushed it quite a bit in Fallout3, Farcry 2, L4D, CSS, Crysis etc. Right now I'm using a HP DV7 with a low spec CPU and 4650 GPU and it eclipses the 3670 I had. Sorry to be the bearer of bad news for ya.

 

Just to add, I have had my M1530 for over 2 years and its been overclocked (GPU) since day one. No problems whatsoever - it scores just under 6100 3d06 on the default res. The one thing you guys are missing is the overclocking ability of the 8600GT, it really does fly. No need IMO to upgrade from that to the newer ATI's if your overclocking.

And I've never used a cooling pad on it!. Missus is playing Overlord II on it right now and she has no complaints .

 

When did you have that XPS 16?

Anyway, relax. I didn't say the performance was equal, I said the 3670 isn't that bad. Besides, notebookcheck's numbers aren't known as being reliable at all on this forum.

Now, like I said, with good drivers (and remember, ATI has been known for being slow to provide decent drivers for deskops, notebooks have been closely to ignored) the 3670 can do 20+ fps on Crysis Warhead, gamer settings and resolution of 1024x768. No way in hell if the 4670 can do anywhere close to 40+, like you claim. So sorry, you are wrong. Unless you can provide benchmarks that contradict this.

This is all I'm saying.

 

1024x768 is a pathetic resolution for starters. Secondly I said in MOST cases the 3670 is 50% of the performance of a 4670. Other factors such as CPU limitations can come into play depending on the game tested as well as optimizations in place. ATi's drivers have been stellar for years, I don't know what you're talking about. Notebookcheck's numbers hold up to other benchmarks I've looked at for 4670 as well, go have a look at the thread I made, I posted some MSI 4670 (512 MB) numbers. Like I said, let the 4670 XPS 16 arrive and I'll gladly prove my point.

Oh and to answer your question, I had the SXPS 16 only a few short months ago and know exactly how well the 3670 performs. That's why I'm saying..wait and see... Not every benchmark will show a 200% disparity but quite a few will reach that point or get close enough. Once you admit the 3670 really is lacking in performance and you still can't find the 4670 where you live, shoot me a PM and I may help you get one.

 

1024x768. 1920x1080. I don't think it will matter much .
The XPS 16 had a decent CPU (2.6 ghz t9550) and achieved very impressive results on WOW and Crysis Warhead. No way in hell 4670 performs 100% better in those games (that's in area of the top end mobile GPU's). More like around 40%-50% better. And I think this is the case for most games with 3670 vs 4670 as well.

 

I don't really see what people are saying about the 3670 being a low end card now to be much use, yes on the scale of graphics it is at the lower end seeing as the nvidia 280 sli completely decimates it so it's closer to the bottom end which would be integrated gpu's. What really makes the comparison though is that it is still a hell of a lot better than what most people have in their computers (which would be an intel HD4500 or something similar), most computers being sold have integrated graphics which compared to the hd3670 are absoute snails when it comes to frame rate crunching

 

Depends, if those games are CPU dependent then the GPU won't be as big of a factor. If they aren't, then the 4670 will slap around the 3670. I haven't looked into Warhead benchmarks yet to determine if is optimized any differently than Crysis yet. I do know Crysis is a major GPU hog and in that game, the 3670 gives about 50% of the performance. We've got 4670 owners with the T9500 so I'm sure they will gladly oblige with benchmarks. I've got a 2.53 ghz chip in mine so I'll benchmark against people that have similar configs with a 3670. I'm also certain if you could, you would jump on the 4670 and not look back, am I right?

 

Agreed on all points but when speaking of discrete chips, the 3670 falls near the low end these days. The 4670 is also very middle of the road, it's a budget chip in ATi's lineup. I'm sure the Sager and Asus GTX 260m guys would have a laugh at this thread because that card destroys both the 4670 and 3670 handily.

 

Lol, for sure, but hey; when you bought a "slim" notebook, you knew you had to sacrifice performance for looks and portability ( dont know how portable the sxps 16 is, since its a bit wide... )

as far as im concerned, as long as you can play games you like at decent FPS, it doesn't matter if its the 4670 or the 3670, or even the 8600gt

for those looking for high end graphics, im afraid your with the wrong models to start with ( well at least with the xps slim line, because the m1730 is a beast )

so as the title of this thread states " 8600gt or ATI 4670" the truth is that the ATI card wins, because its a new card, but the 3670 and the 8600gt still work really well at gaming, not to mention the 8600gt overclocks like a charm. ( dont know about the ATI card, never used it tbh..)

hehe you dropped me to the floor with this one, and no *looks at m1530*... its still an awesome notebook

 

well, GTX260M destroys them in consumption too. it eats 60W, i don't think i want a laptop with that kind of consumption..

 

How do you overclock your gpu on the m1530. Can you give me some pointers?

 

Which is it? 'Not much' better or 50% better? You seem a little confused..

 

btw if u look at the benchmarks, u can clearly see, that 4670 performs OVER 100% better, than the 3670. it just blows it away.

 

No, you are. Joker said it performed 50% Less then the 4670.

Scorpion, those benchmarks are outdated.

 

sure they are bro, sure they are.
let's see then.. 3DMark Vantage.
3670: 1277
4670: 3062
and sure, these tests are all outdated too: (if u see, same driver version was used for both of them, so ur ridiculous state, that it's just all the drivers fail here too)


you can't argue with the facts, bro. 4670 blows away the 3670, with an average of 2X performance. thats a fact bro, just admit it.

 

Relax there, street gangster. First of all, I wouldn't trust those AMD graphs, as they are aimed at marketing their new 4670 card.

Second, just take a 3670, put in new drivers (I don't know why, but the card simply performs far better with new catalyst drivers then it does with the old ones), and you will be able to play Crysis Warhead just fine. Hell, you are saying a 4670 will manage Crysis Warhead on gamer settings, 1024x768 at 40-45 fps. And WOW, maxed out 1920x1080 on 80-120 fps.

 

Guys, Kirov is in denial, I've seen guys like him many times. There's no point in presenting facts because he'll keep making excuses. Yes yes the 3670 actually performs FASTER than 4670 Kirov because of new drivers..cuz you know..

Do you believe in magic?
Yeah, do you believe in magic, yeah
Believe in the magic of ATi Catalysts
Believe in the magic that can set you free
aaahhh, im talkin bout the magic
Do you believe in magic, yeah
Do you believe like Kirov believes
Now do you believe like Kirov believes in magic
Do you believe like Kirov believes
Do you believe like Kirov believes

 

There's no reason to be jerks, he has valid points. He may or may not be incorrect, but he's right about it being an AMD chart. You go "ohh wow, the 3670 was powerful, and look at how much of a boost the 4670 is!!"
I'm sure there's two people on this forum who have similarly configured systems, one with a 3670 and one with a 4670 that could run a benchmark with the same drivers installed. That would settle things one way or another.
If you're so sure of your position, you should support this.

 

Maybe you haven't read carefully enough but that's exactly the position I've taken. But the facts are far too evident--independent benchmarks confirm exactly what AMD put out, that in a large number of cases, the 4670 is 200% the performance of a 3670. Chalking it up to drivers is downright ridiculous and an assertion as such deserves ridicule, especially when he completely dismissed the fact that number of shaders and transistors essentially doubled and that it's a generational leap over the 3xxx series.

 

What exactly are the facts? That 4670 runs 2x as fast as 3670 NO MATTER WHAT??! I'm not trying to be a jerk here.
Joker, first of all, relax. I never denied that 4670 was significantly faster then 3670.

Second, I have pretty much proven my point with the Crysis Warhead and WOW testing. New drivers DO make it far faster (and ATI has been known for being slow with decent drivers).

 

The facts are the hard benchmarks available and you choose to ignore it and chalk it up to old drivers. You keep making a circular argument about Crysis Warhead fully knowing the 4670 configuration cannot be tested yet given the fact that none of us has received it yet. However, what we can do is point to 3rd party benchmarks of the 3670 vs 4670 that do show a large disparity in performance, up to 200%. Furthermore, you're given the reason behind the performance delta between the two GPUs (large hardware differences) and again you brush it off as meaningless. You choose to stick your head in the sand and ignore all reason and that merits ridicule and lack of credibility.

As for Crysis Warhead and WOW, you haven't proven anything. Driver revisions tend to optimize for certain games, if you knew anything about graphics card technology you'd be aware of this simple fact. Not to mention this optimization would also apply to a 4670 and whatever difference there was between a 3 series and 4 series in those games would still hold.

 

While I agree that the increased shaders does make a difference, the whole generational thing doesn't matter really. For example, 7600GT and an 8600GT perform almost identically.
And notebookcheck is a really, really unreliable source of benchmarks. They can usually give you a slight idea of which card is better, but should never be taken more than empirically. Unless I missed some other third party benchmarks in this topic. If I had to guess, I'd say it's probably about 30% better. But that's just my guess based on specifications published by the manufacturer, and not based on real world performance. So my guess too should be taken empirically.

 

You will never see a 200% increase in performance across the board, I haven't and wouldn't make such a claim. However, in many cases you will see exactly that. This could be at a certain resolution and with or without AA/AF enabled. ATi video cards are highly dependent on the number of shaders, they have been since R600. The performance generally scales along the number of shader and clock increase with their cards. nVidia on the other hand works completely differently since they traditionally have less shader pipelines but more ROPs. Thus my argument that a generational jump for ATi, especially one involving more than 2x increase in shaders does in fact create a large performance delta, especially in modern games that make heavy use of shaders.

I currently own a 4650 based laptop and the only difference between it and the 4670 is clock speed. My previous XPS 16 had a 3670 in it with a 2.4 Ghz chip and in Left 4 Dead alone, I've seen a substantial increase in performance. Keep in mind Left 4 Dead is a heavily CPU dependent game and that this HP has a weaker CPU (2.0 GHz T6400 vs 2.4 Ghz P8600). Anecdotal testing shows that the 4650 coupled to a weaker CPU is still providing 150-160% the performance of the 3670 configuration I had before. For example, in the map Dead Air, when the airplane comes crashing down at 1600x900 max settings, the 3670 used to drop down to the 30s (approximately) while the 4650 stays pegged at 60+. During heavy gameplay when the zombies start attacking, the 3670 often struggled and would drop down again to the 30-40 fps range, conversely the 4650 stays at about 50-60 fps--not a 200% gain but like I said, it's a weaker CPU and the game is sensitive to CPU increases.

This will easily be settled once those of us waiting on a 4670 receive our systems so we can do a true head-to-head benchmark. However, I surmise your 30% estimation is way off.

 

Well, you've got more experience with it than me, so I would expect your estimate to be better judged than mine.
At the same time, I really do not think that the 3670 -> 4670 gives a 200% boost in performance across the board. I guess we'll have to wait for some good benches.

 

Again I'll reiterate, I do not think it will be 200% across the board either. I never made such a claim.

 

ok, then just keep waiting an let the facts speak.
those who expect only 30% better performance: i'm pretty sure u'll be surprised soon. there could be cases, when there's only that difference, but in most of the cases, generally, you can expect far more.
u could see it right now, if u check any benchmarks of 3670 and 4650 (which is a bit even slower, than 4670).

 

The graphics isn't the only thing that determines frame rate, remember the HP DV5T with the 9600m gt? For it's money it was a great gaming machine and could easily pull off 30fps in crysis, yet other PC's with that card sturggle to get ten. With the new graphics dell may have redesigned the mobo, maybe for the worse if it turns out to be a pile of crap. I'm just saying that raw numbers don't constitute to a good gaming PC.

Another thing to mention........ learn how to use percentages properly! Seriously, people say there could be a 200% increase in performance (highly unlikely) when they mean to say 100% and other people think they actually mean 200% rather than it being a mistake. This is one reason why we have so many arguements here. Either learn how to use percentages, post in multiples (ie 2x) or don't post performance gains at all

 

The way I've posted percentages aren't wrong at all. It just comes down to how you interpret it but the math is solid. For example, you could say the 3670 is 50% slower or that the 4670 is 200% faster, they both mean the same thing. Furthermore, I don't think you will see much of a motherboard revision in the new SPXPS 16, just a simple swap.

 

Wrong, when you mean to say a 2x increase in speed, it would be 100% faster

 

No it's not wrong at all. If you reference the 3670 performance at 100%, a two fold increase is 200%. The math is not wrong but I think your understanding of it is. There's no point in arguing if you can't grasp this.

 

I think we're both right to some extent, I have come to the conclusion that we should just ditch percentages