AVADirect's W860CU: Mobility Radeon HD 5870 Vs. GeForce GTX 285M

Page 2 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Status
Not open for further replies.

notty22

Distinguished
I'm getting used to ATI's gpu's under performing and having their fanboys looking for scapegoats. ATI is now [strike]re-branding[/strike](naming scheme,scam), lol. If you have a ATI gpu in your notebook, get the extended warranty.
 
Yup, not surprised. Hey, if you want to game on a laptop you have to pay more to get the same performance as a cheaper desktop alternative. Kind of wish nVidia would stop pushing the G92, but I was a bit surprised it was able to keep up so well with the mobile 5870 despite newer parts more advanced process and technology. Then again, I guess DX11 transistors eat up a fair bit of power which may be why the AMD part isn't clocked higher. Still, I'm sure it costs AMD less to make a mobile 5870 than it costs nVidia to make a mobile 285M.
 

saliwok

Distinguished
Aug 14, 2009
13
0
18,510
Excellent article. Thanks so much for doing this.
I'm currently in the market for a gaming laptop (because I don't have room for a desktop) and this is just what I was looking for. I really hope that you are able to do the same thing when we finally see a dual Mobility HD5870 Crossfire machine on the market that you can compare with a dual GTX285 SLI. Then again, it might be cheaper for me to buy a bigger house and get a desktop.
 

dragonsqrrl

Distinguished
Nov 19, 2009
1,280
0
19,290
[citation][nom]juliom[/nom]You're mistaken. The 4850 performs roughly on par with the GTS 250, not the 4870 which is way faster. The 4870 performs as a GTX 260 with 216 cores and the 4890 like a GTX 275.[/citation]
Mistaken? You're correcting a statement that was never made. I never suggested the GTS 250 performs on par with the HD4870. On the contrary, I actually said "the HD5770 performs roughly on par with the HD4870". I also said "on average these two cards perform just a few percentage points faster then the GTS 250", which in most cases is entirely true.

Correct me if I'm wrong, but you seem to be suggesting that there's a huge performance gap between the GTS 250 and the GTX 260 c216 (and therefore the HD4870) as well as other last gen high end cards, which is not true in most cases. Try double checking any reasonable number of benchmarks for popular games (Crysis Warhead, Far Cry 2, MW 2, etc...). Even when comparing to the GTX 285, there's certainly a large performance gap between it and the GTS 250, but it also isn't as large of a gap as you might expect from this sort of a comparison. It's certainly no where near 2x the performance, like one might expect.

Given a good spread of benchmarks, the GTS 250 performs somewhere between the HD4850 and HD4870. And yes, on average it's just a few percentage points slower then the HD4870.

This is all a very small portion of my previous comment and completely beside the point I was trying to make, regarding the performance and power consumption of the Mobility HD5870... just to keep this all in perspective.
 

amdgamer666

Distinguished
Oct 24, 2008
101
0
18,680
[citation][nom]demomanisbest[/nom]This doesn't make much sense to me... if a 5870M chip = roughly a 5770 desktop chip and a 285M = roughly an 8800gts.. why is it not completely spanking it? we all know 5770>8800.. by a rather large margin! what could be the cause of this?[/citation]

I'm almost positive the 128 bit bus width is causing a(n unfortunate) bottleneck.

We all know that the desktop 5770 also has a memory bottleneck.

Desktop 5770: 1.2 GHz GDDR5 x 128 bit = 76.8 GB/s; 1360 gigaflops
Mobility 5870: 1 GHz GDDR5 x 128 bit = 64 GB/s; 1120 gigaflops

1120/1360 = ~82.4% raw power
64/76.8 = ~83.3% bandwidth

So there u have it. Now it'd be foolish of me to forget to mention that the gtx285 is also memory bottlenecked due to the use of GDDR3.

1.02 GHz GDDR3 x 256 bit = 65.28 GB/s

So it only has a little bit more bandwidth it's something. I'm not sure if that can quite account for the results but it's something to keep in mind.
 

dmccarron

Distinguished
Feb 3, 2009
37
0
18,530
[citation][nom]killerclick[/nom]"Gaming notebooks" are for suckers with too much money. Putting up with slower, ridiculously expensive hardware and smaller screens in order to be able to play in school or... where? Where do you "gaming notebook" guys play your games anyway?[/citation]

"gaming notebook" machines fall in the same class as mobile workstations, just with the Radeon or Geforce instead of the FireGL or Quadro. In many cases the former are selected instead because the Quadro's and FireGL's cost even more and the performance is lower in lieu of rock-solid, red-carpet driver support.

A sizeable fraction of the buyers for these high-end rebranded Clevo/Eurocom clamshells(e.g. the direct Sager resellers in the US, not outfits like Alienware or Voodoo) are folks who can make use of the performance outside of, or at least in addition to, such gaming.

The cost of a powerful, mobile computer of this type with a 17" to 20" screen and all the goodies is on the order of $3000-$5000. A lower-end (i.e. mainstream) graphics option is often unavailable as an option and in any case yields a savings of a couple hundred dollars at most (not an issue on such a high-margin product to begin with).

I've seen a few BS money-sinks (like the Dell WoW-bedecked special-edition XPS a few years ago, for example), but let's not forget that a notebook is not a glorified gaming console, and the 'suckers with too much money' you mentioned I believe are outnumbered by others with salaries commensurate with their abilities whose job and travel schedule easily justifies the expense. Oh, and most of these folks have grown up on video games. Being able to play the occasional title on top of everything else never hurts.
 

gallidorn

Distinguished
Jan 23, 2009
104
0
18,680
Why don't Intel and AMD incorporate hybrid crossfire or SLI to use the integrated graphics resources to provide an additional performance boost?

They do this with desktop systems, so why can't they do this on a mobile platform to make it closer to the similarly labeled desktop model?

Come on guys this should be the obvious direction for improving performance. Especially since operating systems and applications are becoming more multi-thread capable.
 

amdgamer666

Distinguished
Oct 24, 2008
101
0
18,680
[citation][nom]killerclick[/nom]"Gaming notebooks" are for suckers with too much money. Putting up with slower, ridiculously expensive hardware and smaller screens in order to be able to play in school or... where? Where do you "gaming notebook" guys play your games anyway?[/citation]

srsly? Less than a week after I've gotten my new HP Envy with a mobility 5830, I've already taken it to a LAN. Use your brain
 

amdgamer666

Distinguished
Oct 24, 2008
101
0
18,680
oh and one last thing. model numbers are just that: some arbitrary number. The majority of the time the notebook graphics are gonna be a severely crippled desktop part or just based off of a lesser gpu. I don't really mind that either of them aren't that similar to their desktop counterparts of the same number. Just get used to it. What really bothers me is nvidia rebadging their desktop parts. All you have to do is continue to sell your older parts, you don't have to rebadge!
 

hamitaltintop22

Distinguished
Aug 13, 2008
98
0
18,630
[citation][nom]dragonsqrrl[/nom]Mistaken? You're correcting a statement that was never made. I never suggested the GTS 250 performs on par with the HD4870. On the contrary, I actually said "the HD5770 performs roughly on par with the HD4870". I also said "on average these two cards perform just a few percentage points faster then the GTS 250", which in most cases is entirely true.Correct me if I'm wrong, but you seem to be suggesting that there's a huge performance gap between the GTS 250 and the GTX 260 c216 (and therefore the HD4870) as well as other last gen high end cards, which is not true in most cases. Try double checking any reasonable number of benchmarks for popular games (Crysis Warhead, Far Cry 2, MW 2, etc...). Even when comparing to the GTX 285, there's certainly a large performance gap between it and the GTS 250, but it also isn't as large of a gap as you might expect from this sort of a comparison. It's certainly no where near 2x the performance, like one might expect. Given a good spread of benchmarks, the GTS 250 performs somewhere between the HD4850 and HD4870. And yes, on average it's just a few percentage points slower then the HD4870.This is all a very small portion of my previous comment and completely beside the point I was trying to make, regarding the performance and power consumption of the Mobility HD5870... just to keep this all in perspective.[/citation]

I am sorry for being tedious, but hd 4870/gtx 260 and to a degree the hd 5770 are faster than gts 250 by more than a few percentage points. If you said a few fps then it might be true. Think about it, sometimes a small fps difference can mean a much larger % difference.
Example: I am looking at Toms' 2010 charts and benchmarks for Battlefield Bad Company 2 (1920x1200, 0AA, 0AF, High Quality, DX10). The hd 5770 gets 36.2, gtx 260 gets 35.3 and gts 250 gets 32 fps. The fps difference is only about 4 between the hd 5770 and the gts 250, but the hd 5770 is 12.5% faster than the gts 250. Of course, this is only one game at one setting with outdated drivers (ATI cards would probably benefit the most with a driver update because it is a much newer generation).

All I want to say is that there are distinct fps performance categories. The HD 4870, gtx 260 and to a certain degree the HD 5770 are in a category which is significantly faster than the category below it(Hd 5750, hd 4850, gts 250, 9800 gtx +, hd 4770 and 8800 gts 512). While I don't like putting the hd 5770 in the same category as the hd 4870 and the gtx 260, due to its performance reducing memory bandwidth, the hd 5770 is much closer to the hd 4870 than the gts 250. And there is a significant performance difference (whether in frame count or %) between the hd 4870 and the gts 250.

NOTE: I own a 8800gt, and thus my comment is not influenced by the need to justify my ownership of any of the cards stated.
 

killerclick

Distinguished
Jan 13, 2010
1,563
0
19,790
[citation][nom]dmccarron[/nom]Being able to play the occasional title on top of everything else never hurts.[/citation]

Hunched over a puny 17" screen at an airport lounge and playing Modern Warfare 2 on a $3000 eight pound overheating laptop in my lap for 30 minutes until the battery runs out isn't my idea of top-notch computer entertainment. I don't know, I just don't see gaming as something worth doing 'on-the-go', much less paying so much for an inferior experience (compared to the desktop). Sure there are people who need serious number crunching in a portable form, but look at what this kid replied:

[citation][nom]amdgamer666[/nom]srsly? Less than a week after I've gotten my new HP Envy with a mobility 5830, I've already taken it to a LAN. Use your brain[/citation]

And a lot of other gaming notebook buyers are like that, just kids with money to burn. Fine, it's their money (or their parents' more likely) but they're still suckers for paying so much, especially if they only play on a laptop. So yeah, I don't get it...
 
G

Guest

Guest
the gts 250 has absolutely no ground even close to the 5770 in any game I have except borderlands. I own 2 xfx black 260's a xfx 5870, a 5770, a 4890 and 5 4850's that i won. even a single ati 4850 with 512mb can beat the gts 250 in all of my fps games except borderlands by 1-6fps not a whole lot but hey its almost 3 yrs old now. stepping up to a gtx 260 feels like double the performance its not even close especially in fallout 3 and battlefield bad company 2 the 5770 is only on par with the gtx260 if its bone stock clocked. the guy was right about these performance numbers looking fishy like someone got paid by nvidia to not look so bad. the 5770 is clearly a better card than the 250 by my own testing on three platforms from a q9450,i5-670,i7-860,i7-920 oced and the gains were from 16%-35% on all the intel platforms and on my amd machine a oced phenom 940 it showed smaller gains 8%-20% on all fps's except borderlands. thers no reason why the 5870m shouldnt of beaten the gtx285m by atleast 15% in every game unless the cpu was bottlenecking the setup wich is a possibility bec these numbers looked alot like all the amd processor gaming benchmarks out there with 4890-gtx285 and above multiple card configurations. the tdp differences made absolutely no sense.
 

pwnagehobo

Distinguished
May 8, 2010
20
0
18,510
[citation][nom]GairB3ar[/nom]I own 2 xfx black 260's a xfx 5870, a 5770, a 4890 and 5 4850's that i won.[/citation]
[citation][nom]GairB3ar[/nam]...the 5770 is clearly a better card than the 250 by my own testing on three platforms from a q9450,i5-670,i7-860,i7-920 ...[/citation]
And I thought I owned too much (and spent too much money on) PC hardware... And you only mentioned the recentish stuff. :p

Respect++
 

boletus

Distinguished
Mar 19, 2010
69
0
18,630
If I had the dough, I would get one. Why? Because my job sometimes forces me to spend a week in a motel, and I HATE television. So I get bored out of my mind and wish I could play even a few older games. But I may break down and get $400 Acer or something, just so I can surf the net and play solitaire.
 

amdgamer666

Distinguished
Oct 24, 2008
101
0
18,680
[citation][nom]killerclick[/nom]Hunched over a puny 17" screen at an airport lounge and playing Modern Warfare 2 on a $3000 eight pound overheating laptop in my lap for 30 minutes until the battery runs out isn't my idea of top-notch computer entertainment. I don't know, I just don't see gaming as something worth doing 'on-the-go', much less paying so much for an inferior experience (compared to the desktop). Sure there are people who need serious number crunching in a portable form, but look at what this kid replied:And a lot of other gaming notebook buyers are like that, just kids with money to burn. Fine, it's their money (or their parents' more likely) but they're still suckers for paying so much, especially if they only play on a laptop. So yeah, I don't get it...[/citation]

1) Even a "puny" 15" is plenty when you're only a foot and a half from the screen. It's not my fault that you have shitty vision and can't see a damn thing, old man.
2) I paid only ~$1030 after tax.
3) I didn't have "money to burn", I saw an excellent deal on an excellent machine and decided that it would be the perfect upgrade. This wasn't some luxury item or neat toy to some rich, spoiled douche.
4) Not only does my laptop not overheat, it runs incredibly cool.
5) It has 2-3 hours of battery life; not the greatest but not bad at all. I haven't even tried to game without plugging in because there's no point, we both know that. Though I imagine I could get at least an hour
6) I don't understand why you don't understand why gaming isn't worth doing on the go. The game doesn't change just because a person is away from an outlet or if the computer isn't stationary.
7) I'm not fatal1ty, I don't have to have flawless framerates all the damn time. It's just games, the point is to have fun and I can now do that on the go :)
 

stonedatheist

Distinguished
May 13, 2010
97
0
18,630
[citation][nom]gallidorn[/nom]Why don't Intel and AMD incorporate hybrid crossfire or SLI to use the integrated graphics resources to provide an additional performance boost?They do this with desktop systems, so why can't they do this on a mobile platform to make it closer to the similarly labeled desktop model?Come on guys this should be the obvious direction for improving performance. Especially since operating systems and applications are becoming more multi-thread capable.[/citation]

The problem is that the integrated graphics have absolutely dreadful performance in comparison. The only reason they do that in the first place is because it's actually meaningful when they're used to improve the performance of a low end card. An hd3200 is almost as powerful as a 3450 so enabling hybrid crossfire in that scenario almost doubles performance.
 

dmccarron

Distinguished
Feb 3, 2009
37
0
18,530
[citation][nom]killerclick[/nom]Hunched over a puny 17" screen at an airport lounge and playing Modern Warfare 2 on a $3000 eight pound overheating laptop in my lap for 30 minutes until the battery runs out isn't my idea of top-notch computer entertainment. I don't know, I just don't see gaming as something worth doing 'on-the-go', much less paying so much for an inferior experience (compared to the desktop). Sure there are people who need serious number crunching in a portable form, but look at what this kid replied:And a lot of other gaming notebook buyers are like that, just kids with money to burn. Fine, it's their money (or their parents' more likely) but they're still suckers for paying so much, especially if they only play on a laptop. So yeah, I don't get it...[/citation]

Personally I have never owned a high-end notebook yet. (Got a dozen or so expensive, decked-out desktops from over the years that are more show-pieces than really used as computers now). I agree that the high-end performance notebook parts tend to be overpriced for what you get. But then again, there are people who turn pickup trucks into racing vehicles (why not get a car for the job?) or spend throw a lot of money to trendy clothes (why not just throw it at a dancer in a cathouse?). An enthusiast in any arena has a hard time justifying their desires. One could of course take it to the next level as well by claiming that PC gaming enthusiasts who blow $3500 on a top-end desktop compared to a good console and a giant LED TV. Of course, the advantages are the additional versatility of the PC, higher possible resolutions, and lately the real prospects of native multi-headed gaming, but surely there's a die-hard console fanboy who refuses to acknowledge these in light of the costs.

Anyway, I did get a chance to play with the Turion-based Clevo machine with the 20" screen a couple of years ago for an hour or two to see what it looked like. The current 18" models are rather attractive too, if you like blue lights everywhere.

All my monitors are certainly far above 17" or 20" (whose aren't, these days?), but I never really got the feeling that a 17" screen on a notebook was too small to enjoy. I'll take the 20" screen any day, however, to be sure. I'm not sure battery life is much of a problem for most of these guys either since they're not doing anything too far from an outlet - no Crysis out in the woods, I bet.

In any case, if external PCI-express-based video cards (like the attempt with the Acer netbooks) ever catch on, then there's the time for all of us to start thinking seriously about mobile gaming.
 

babachoo

Distinguished
Nov 17, 2009
271
0
18,790
For the desktop comparison, they should have used a slower CPU since most i7 laptops use the I7 720. They also should have used a Radeon HD 5770 or 5750, since the mobility version of the 5870 is really just an underclocked Desktop 5770.
 

Crashman

Polypheme
Former Staff
[citation][nom]babachoo[/nom]For the desktop comparison, they should have used a slower CPU since most i7 laptops use the I7 720. They also should have used a Radeon HD 5770 or 5750, since the mobility version of the 5870 is really just an underclocked Desktop 5770.[/citation]Most i7 notebooks might use the 720, but THESE used the 820, which means THESE had the same clock speed at one or two threads as the 920. What you forgot to say is that Tom's should have used a slower i7 desktop CPU (one that doesn't exist) AND told the builder to use slower notebook CPU's (that would have made their own product look worse) simply to satisfy your curiosity about some other notebook hardware that wasn't even intended to be part of the test.

I personally think AVADirect got it right.
 

babachoo

Distinguished
Nov 17, 2009
271
0
18,790


Huh? I tried to make some sense out of your post, but I just can't do it.

#1 - The 820 is still 1.73 GHz - MUCH slower than a desktop i7 920.
#2 - They DO make slower i7 desktop cpu's, for instance the i7 860S which clocks at 2.53.
#3 - My point is that by comparing these laptops to a far superior desktop, rather than one with comparable specs, Tom's didn't really think through their benchmark comparison. Even keeping the comparably blazing fast i7 920, they could have at least used a HD 5770 or 5750, for reasons I've already stated.

Make sense now?
 

Crashman

Polypheme
Former Staff


No, look at the Turbo Mode speeds, since Turbo Mode was enabled for ALL systems. The 820 and 920 run at the same speed in games. The 820 runs faster than the 860 in games. It's all about turbo mode when you're using less than four cores.
 

thejames

Distinguished
Mar 3, 2010
4
0
18,510
I cannot believe these clowns at AVA Direct are still smoozing up industry bigwigs while screwing over any little guy they can.
http://avadirect-ripoff.blogspot.com/
 
Status
Not open for further replies.