GTX 460m vs ATI Radeon 5870

Wouldn't worry too much about your Furmark temps. They look well within spec. Sure cooler temps are nice, but again, it's well within the norm...

 

I'd still prefer to see mid 70s

 

With a single fan cooling both ? I would say your temps are better then most that have 2 fans lol...

Be content with your setup and leave it alone ..

 

Never!!
With manual fan control I can now land a 940XM there and OC' it to the i7-975 levels...maybe...

 

Thats about the only upgrade i may do on this machine, 940xm ... We shall see though..

 

nah, was not looking for you to do it this go round. was looking for all these other new owners looking for names...

 

I think in time we will see more of it but its far to early and only a few people have received them but theres more every day saying they have got them or just ordered etc...

 

i was just going by the title of the thread..haha
thought this was the new thunder dome...just not yet.

joker already posted a run for the x7200...just waiting for that battle royal to really get kicked off.

 

Yeah thats far more exciting to me as well..

My machine is just good fun and those machines are in a whole nother league ..

I guess with Sager only ordering 6 SLI GPUs it might be awhile before we get to see some real numbers ...

 

Hey Quad,

Did Asus fix all the bugs in the JW that faced the g73Jh when it first came out?
you know, the BSODs, keyboard lag, VBIOS, LEDs OFF, freezes, (some 5870s had cooling issues...)
the thread was 350+ pages long lol.

 

There are currently no issues to report with the JW..

No GSOB,BSOD,VBIOS etc etc etc ... Temps are amazing on the 460m right out of the box ...

 

Hahaha! I also have a lot of games and don't have problems launching games! Long live ATi! XD Seriously, where do you come with this stuff? What is this crazy talk about making file directories to make games work?

Oh well. Seriously, both companies offer similarly priced/performance products. There is no much point in arguing which one is better, if they perform rather close to each other.

Go with what you want. I am perfectly happy with my ATi card, and I would not trade it in for a 460m. Why? It's the same freaking thing. Waste of time and money. Got a 460m? Then no point in getting a current gen ATi card.

 

hey guys! i just run the unigine.beta driver, 810/1620/1450

 

Thanks Spirit32 saved me the trouble of running it ..

 

The Ryzeki speaks the truth. (Except for the "ATi" part. It's AMD now, dammit)

I would categorize myself as someone who would, all other things being equal, would get an nVidia-equiped laptop instead of an AMD one, but I'm perfectly happy with my 5870 and I'm not looking back.

In the end, both are amazing cards.


That being said, I would love to sidegrade (is that a word?) to a 460m just to see if I could.

 

Sure, lets fight I OC'd mine only to 860/1030. Im not even among the strongest HD5870m haha. I wish I could OC my memory like you so I can trully shine.



Link to my pic


Looks like even in DX11 I can still put up a fight eh?

Me loves my temps

 

Wow. That's rather incredible really considering it's DX11. Months ago johnksss was telling me the minFPS in Nvidia cards are inheritently superior to ATI, I tried to tell him basing his assumptions off of intuition rather then empirical analysis was a mistake.

Perhaps he is right though, would like to see more great benchmark comparisons like this.

 

Yeah, and I would love to compare benches/games too. Both cards and notebooks are amazing on their own.

As for the framerate having higher min fps in nvidia cards, I don't know. We will find with more benchmarks as the last ATi/AMD card I had before my HD5870m was a X700 64MB.

The only comparison I have available is with my previous GTX260m, where in some cases, the HIGHEST fps of my GTX260m are the MIN of my HD5870 (one of them being metro 2033)

 

Well in 285 vs 4870 the higher min FPS went to the 4870, as well as general stability in FPS of the benchmarks Don and Pman tried. I tried to argue with john he was making a mistake suggesting that but he wouldn't listen, hope future benchmarks clarify the cards are rather equal in prowess in a pragmatic sense. I guess I'm tired of hearing , AMD THE BESTEST, NVIDIA THE BESTEST EVAR! Really it's brand preference as the cards really draw equal in realistic use.

 

some more comparison. there isn't really that much difference!

 

3 things.
ask quad
ask speedy gonazales
ask baby hemi
all of which used both types of cards.

 

My impression was that in general, with current HD5xxx vs 400 and previous HD4xxx and 200 series, Nvidia generally had higher Max FPS, but had massive dips into low FPS causing stuttering even though it could attain incredible max FPS numbers
- One of the things AMD did well from reviews I've read is consistency. Despite lower framerate in games, games performed fluidly without stuttering etc because of consistent framerate.

Can disagree with me, no matter to me, would just be the reviews I've read. IMO that's where the other value in AMD comes in, the average FPS and minimum FPS.

Just want to mention considering the 460M are just now starting to be released, with only one laptop having it and maybe few more in October. AMD HD6xxx launch is in October... Who wants to bet that HD6xxx mobile will be released soon after and will still cost less than the 460M? Just saying, I base a lot of my judgement on hardware by price/performance.

 

Ryzeki what do you get in Uniengine heaven benchmark in DX9?

 

My thoughts exactly. Having owned 280's and 4870's (M17X) I noticed that while the Green's had higher max FPS, the Red's usually had higher low FPS in most games I played.

 

Here you go mate, enjoy.

 

DUDE, c'mon. Stop posting such massive pics. Use thumbnail or something, this is ridiculous.

 

Hahaha, come on, lets make the low resolution users suffer!

jk jk, ill post thumbnails next time onwards

 

ryzeki here is my uniengine score in DX9 on my XPS M1730

http://img687.imageshack.us/img687/712/unienginedx938ghz.jpg

The CPU flag there is a little funny since I don´t run my CPU at 4.3 Ghz, it runs at 3.8Ghz though apparently Uniengine can´t detect this.

 

Well, a pair of OC'd GTX8800 are still a force to be reckoned with. I still managed to get 72% of your score with a single card

And well, 3.8ghz CPU is monstrous too! Man, I wish I had the money to get a CF 5870m, but at this rate, I rather save and see if I can upgrade to 6870m's.

I shall fight back when I can achieve higher clocks. I will have to raise the volts a bit and get 900+ core.

 

ryzeki yes definitely, the 5870m is a very good GPU. Crossfire of those would be something

 

I've been following diffrent threads concerning this issue, and I've also posted in another thread on this forum.

I've seen that they excel somewhat diffrent due to the games they are tested on. (Ati comes out as the winner in most thou) And this far I havent yet seen any "real" comparsion in SC2. (Netbook check states the Fps for the 460)

From what I've gathered there are only but a few who have tried out both the 5870 and the 460, my concerns regarding these cards, comes down the the "simple" question:

Which one will run SC2 the smoothest?

(I migth add that I will be buying either the Asus G73Jh or Jw)

 

With these settings, in campaign about 40-55 FPS. And in matches, 1 v 1 about 35-45 FPS. Some of them I lower down, like why do I need Depth Of Field in a strategy game? Or why do I care about transparent shadow? I don't see improvements with appearance them on, but the performance hit is significant. No idea about 460M, don't care either, this is good enough for me. G73jh HD5870M, 800/1125 O/C, GPU temps around 77C.

 

Hey guys, I splurged and bought the new 460m for my clevo w860cu. Anyway, I read a few pages back that Ken from Gentech overclocked a 460m and got about a 8000 score for 3dmark vantage. I just overclocked mine running on stock voltage and a terrible driver (the only one that currently works on the clevo/sager's) and got 8261 GPU score making a 10100 total score! Not too bad considering that my 280m would only get around 7500 total score or so....anyway....I haven't seen much on any sager/clevo laptops running the 460 so I thought I'd post my findings.

 

Man this sean guy is an idiot...
Thank god he got banned.

 

I'm a lil biased against ATI right now since software/driver issues are preventing my 5650 from running SC2 smoothly on anything other than low. Multiple Pylon slowdown prevents me from using 'high' models and map slowdown (like Metolopolis) prevent me from running the game at anything other than low shader detail. My next gaming notebook will have an nvidia GPU for sure.

 

Anandtech has compared the new G73 with the GTX 460M with the 5870 in the older G73.

 

Are these the clocks you use say if you played Crysis Warhead for 4 hours straight? Or are these your benchmark GPU scores, I don't care about the overall since Vantage is idiotic by adding something done by your GPU to your CPU score and PhysX is terrible on CPU...

Here are HD5870M

My personal stock setting I use for gaming and everyday at all times for 3D related, also ran Furmark 4 hours with this (82C): GPU 8624


Benchmark scores: GPU 9121

Your disappointment is your own doing, not AMD. AMD made no claims the HD5650M would run SC2 on high settings, and actually, no one did.

Your bias is unfounded and baseless. The HD5600M/HD5700M are optimized for laptops for GPU switching, low power, with extended battery life while providing GPU accelerated video output and encoding and some HD features. It's good enough for casual gaming on some DX10/DX9 games, but don't expect 1920*1080 on high settings.

For nearly 2 years, every forum has had recommendations for minimum specification for SC2 on high/ultra settings. Blizzard for over a year stated a GTX 8800 desktop card. It's been said repeatedly that SC2 scales very well but on High/Ultra is difficult. Just look at everything it uses, transparent shadowing, HDR on every surface, Ambient Occlusion to the max, Depth of Field, ridiculous amount of texture detail etc. NBR has numerous threads in which they discuss SC2 performance and where you could have asked can such GPU run a game at these settings and NBR member could have given you feedback and recommendations.

I don't understand how you expected the HD5650M would run these settings. Even a Nvidia 330M plays SC2 on low settings, the lowest settings at 1280*800, going to equivalent Nvidia for the cost of a HD5650M makes ZERO difference in your situation.

 

You are using the 10.10's, right? My 5830 doesn't seem to have any issues. I'll double-check it with SCII tonight, though.

 

On my 5870 with the MSI E7405 @ 1.05V I get 840/1100 normal playing clocks and it's about 9066 GPU score in Vantage

 

Ya those are my "gaming clocks" and have not had any issues at all thus far. As far as temps go, I ran Furmark 1920x1080 extreme burning mode for about 45 minutes and the max that I got was like 76C. I'm not one for benching usually, but I'm sure that if I pushed the card more, I could hit around 9000 GPU score. Originally I wasn't thinking that upgrading from a 280m to the 460m would yield such a big difference but I'm quite impressed thus far. I can only imagine 6 months from now the card's full potential when the drivers are more optimized. I'm not trying to say that the 460m is better than the 5870 at all, I'm just saying what I've found thus far. Hopefully the info will bring more light to the 460m's potential.

 

The driver statement is getting really old and it's unfounded. The GF104 has been on the driver list for Nvidia for more than half a year now. What you have now with the latest drivers are likely it. And there is plenty of evidence that shows that when the GTX 480 was released, drivers then and now did not provide the performance boost some had thought.

And if you think Nvidia is going to specifically optimize drivers for 460M I would think you are delusional. Nvidia improvements will be on the GF104 and whatever those are, will be made for 460M. Just like AMD does not optimize drivers specifically for mobile versions. In fact the INF from Desktop and Laptop are nearly identical except with driver/vendor ID support added.

You can read from Nvidia themselves. Most of their driver improvements are with tessellation, SLI and extremely high settings with 4XAA/16AF @ 1920*1200. Not the settings the 460M are capable of.
http://blogs.nvidia.com/ntersect/2010/09/introducing-the-release-260-family-of-drivers.html

 

Your HD5830M is entirely different beast from a HD5650M.

 

Yeah, I'm running the 10.10s. Here's hoping that 10.11 might fix the issues.

Dude, seriously? The two issues I mentioned are known issues and if not for them, I'd be able to run SC2 smoothly with medium settings. I never mentioned anything about running the game in high/ultra settings. The 5650 isn't as low end as you're making it out to be. I did ask in the recommendations thread and people there mentioned that it could run fine in medium - and it did, until I started playing as Protoss. The performance issues totally have to do with software/driver issues.

 

No sorry, that's not the case. The GPU Score for GTX 8800, the recommended minimum from Blizzard if you want performance is around 5,000. The GPU score of the HD5650M is around 2800, it's about HALF of what is recommended minimum. The GPU that is equivalent to the HD5650M is the 335M. Switching to Nvidia would make ZERO difference in your SC2 issue. Simply not good enough. HD5650M is yes that bad for latest games, especially an RTS like SC2. The HD5650M is the MINIMUM to play SC2 on low/medium settings. If others said different, sorry, they were misinformed.

The drivers are great and definitely NOT the issue. I tested this on the Metropolis map vs the AI. I used the highest quality, no optimization, best detail IQ setting possible. I set my Catalyst Settings for my AI to be of High Quality which has been proven to disable optimizations for best appearance, I disabled the surface format optimization which disables the FP16 Bilinear/Trilinear optimization and I set my Texture Quality to High Quality. You can see in the screenshot the performance is as expected and the details are excellent and there aren't any visual flaws. There is NOTHING wrong with AMD drivers for SC2.

Then I built the Pylon limit for 200 units, a Gateway, Vespene Gas Refinery, 40 probes, and waited for AI to come and destroy my pylons. I still get 37 FPS.

 

Wrong! No offense, but the ONLY driver that is working with the 460m while using a sager 8690/ clevo w860cu is the 259.39 which was released on August 12th 2010. That being said, I am not able to upgrade to any newer drivers because they currently do not have any modded INF files available for my laptop. So, my driver is by no means up to date.

Now, to say that drivers don't make a large difference is completely "unfounded." For example, anybody who used a Dox driver clearly saw a large improvement in gaming potential! Search online and see for yourself. Yes, I do realize that they were not officially put out by nvidia, but my point was that the driver made a huge difference in how well your card performs. I was using Dox drivers back when i had my Asus g51 and saw about a 10fps increase in several games....same clocks....same physics x....same everything with the exception of the driver. I also saw about a 800-1000 point increase running 3dmark06.

Oh, and I never assumed (or even hinted) that nvidia would optimize the driver specifically for the 460m. I am fully aware of what they do when releasing a new driver.

Here are my points: drivers do make a difference, or why would they even release newer drivers in the first place? Sometimes they are going to make a large difference with performance, sometimes they don't. Between now and the next 6 months, they will surely release a driver that will improve the performance. ATI/AMD will too.

 

Dox drivers are specifically for performance at a loss on IQ. That's something Nvidia will not do, not to the extent Dox went. And yes I have used his drivers in the past, and overrated, performance boost wasn't much to get excited over.

And again, Dox does not modify drivers. Dox hasn't a clue how to make drivers. Dox does not touch the drivers, all he does is edit the INF for some IQ/Performance settings and his lean towards Performance. If you want the best IQ/Performance, Nvidia official are still better than his.

And yes GF104 was in the driver INF for a long time now, believe it was spotted back in March. GTX 460M is not the ONLY GF104 GPU... So big deal, Nvidia didn't add your vendor/hardware ID until August, doesn't mean drivers haven't already been optimized for GF104 for MONTHS already

 

Anybody with an nvidia 335 or 330 care to test this out? Scrolling over certain areas of the map causes massive slowdown and it's the energy field graphics of the pylons that cause my system to slow down, not the pylons themselves. I'm going to take a wild guess and say that one or both of those issues will be nonexistent on 335/330 systems. I get a nice 30-60fps on most every other map playing as the other races on medium settings. So, again, the 5650 is not as low-end a GPU as you're making it out to be.

 

Absolutely everything on Ultra this time. Nothing custom, Ultra Texture/Ultra Graphics

It must be something with your system, it's not AMD drivers.

 

It's a known issue, at least with the 5650. There was a thread on the bnet forums describing it, but I can't find it anymore.

 

Is that so? Well, I guess that HP accidentally put the wrong card in my laptop, since every modern game I've tried runs on at least medium. There's no way this entry-level card should be able to do that, right? Damn Notebookcheck for lying to me!

Basically, anything that can't play all modern games at 1920x1080 max/ultra settings 4xAA/16xAF with a steady 60FPS sucks for modern games to you, doesn't it?