Mobile GPU rant.

[StrangerGuy]

Dear AMD/Nvidia:

Could you please have at least some kind of standard for naming the models of your mobile GPUs? Like what you do for the desktop parts? It's not acceptable for a 5850m with GDDR3 to be called as such when it's clearly slower than the GDDR5 version by a huge 40%.

BTW why does a high-end laptop GPU gets coupled with slow GDDR3 anyway, and why does a 5650m is allowed such a sizeable GPU clock ranging from 450 to 650MHz? Is all this some kind of bait-and-switch scam to screw us consumers over so we think we got a great GPU but actually is much slower in reality? We never had to worry about all this shady crap when we buy mid-range/high-end desktop cards so why let this shenanigans continue to infect the laptop side?

Thanks and have a nice day.

 

[Dark_Archonis]

Dear AMD/Nvidia:

Could you please have at least some kind of standard for naming the models of your mobile GPUs? Like what you do for the desktop parts? It's not acceptable for a 5850m with GDDR3 to be called as such when it's clearly slower than the GDDR5 version by a huge 40%.

BTW why does a high-end laptop GPU gets coupled with slow GDDR3 anyway, and why does a 5650m is allowed such a sizeable GPU clock ranging from 450 to 650MHz? Is all this some kind of bait-and-switch scam to screw us consumers over so we think we got a great GPU but actually is much slower in reality? We never had to worry about all this shady crap when we buy mid-range/high-end desktop cards so why let this shenanigans continue to infect the laptop side?

Thanks and have a nice day.

I sort of agree with you that mobile cards need to have more clear naming and specs, but here are a few reasons why I think these discrepancies exist:

- cooling
- cost
- reliability
- competitive reasons

Cooling in a notebook is entirely different than in a desktop. The cost of implementing a modern GPU into a notebook (or smaller) form factor is fairly considerable these days.

Some notebook makers request lower clocks for cooling/reliability/cost purposes. Some potentially do it for profit.

Of course I am not discounting more nefarious reasons such as scamming customers or trying to get more profit out of something. I don't know for sure this occurs in every case though.

What my general point is, implementing a modern GPU into a notebook or mobile design is easily more complex and more expensive than implementing it in a desktop form factor. Therefore, much more compromises are made with mobile GPUs than compared to desktop GPUs.

 

[cusideabelincoln]

It wouldn't be so bad if the complete specifications were actually listed, but sadly a lot of info pages do not contain all the important details. Complete, detailed, specifications should be readily available for every model of laptop for the customer to overlook.

 

[Arkadrel]

StrangerGuy

It sounds like your complaining that a mobile part is slower than the non mobile part?
Let me start by saying, its not just limited to the grafics card, its your cpu too.

also Intel,nvidia,amd all do it, their mobile parts are SLOWER than their equally named non mobile parts.

the end 🙂

The reason their slower (than their non mobile counterparts) is because their ment to use less watts than the pc desktop ones, so you have some battery life.

 

[StrangerGuy]

StrangerGuy

It sounds like your complaining that a mobile part is slower than the non mobile part?
Let me start by saying, its not just limited to the grafics card, its your cpu too.

also Intel,nvidia,amd all do it, their mobile parts are SLOWER than their equally named non mobile parts.

the end 🙂

The reason their slower (than their non mobile counterparts) is because their ment to use less watts than the pc desktop ones, so you have some battery life.

Did you even read my post?

 

[jiffylube1024]

Dear AMD/Nvidia:

Could you please have at least some kind of standard for naming the models of your mobile GPUs? Like what you do for the desktop parts? It's not acceptable for a 5850m with GDDR3 to be called as such when it's clearly slower than the GDDR5 version by a huge 40%.

BTW why does a high-end laptop GPU gets coupled with slow GDDR3 anyway, and why does a 5650m is allowed such a sizeable GPU clock ranging from 450 to 650MHz? Is all this some kind of bait-and-switch scam to screw us consumers over so we think we got a great GPU but actually is much slower in reality? We never had to worry about all this shady crap when we buy mid-range/high-end desktop cards so why let this shenanigans continue to infect the laptop side?

Thanks and have a nice day.

Agreed to a point. ATI they need to keep the clockspeeds in check on the 5650's or call it them something different. They make xx30 and xx70 series cards in other segments!! ATI has a good thing going for them with their x4xx, x6xx, x7xx and x8xx lineups, but then they go and do something stupid like rebadge the 4650 HD as the 560V.

All in all I find Nvidia's GPU lineup a bit more confusing, in part because they bothered to make a 3xx series of GPU's (unlike on the desktops), where many of them are just 2xx series cards with higher clockspeeds but the same architecture and node.

-------

Honestly I think that the laptop GPU market is getting clearer (most new laptop GPU's have TDP published for them, and they usually advertise the clockspeeds which is great), but they still need to advertise better the specs, especially the # of pipelines, RAM type and if it uses DDR2, DDR3, GDDR3 or GDDR5.

 

[DaveSimmons]

Agreed, both nv and AMD will sell a part with 5850 level desktop performance but brand it as a mobile 5870.

You have to throw away everything you know about the desktop model numbers, and treat the mobile part numbers as completely unrelated.

I see it as a cynical attempt by nv and AMD to mislead consumers who see "mobile GTX 470" then use Google to find a chart of desktop video card benchmarks, not realizing those benchmarks do not apply.

 

[Schadenfroh]

We need part numbers like this:

M (for Mobile)
D (for Desktop)
P (for professional workstation)

M1 1 (first generation mobile part normalized base)
M1 1.3 (first generation mobile part roughly 30% faster than base)
M3 1 (third generation mobile part)
etc.

Marketing would never allow it to be clear, nor would they agree on how to measure "30% faster."

 

[StrangerGuy]

Agreed, both nv and AMD will sell a part with 5850 level desktop performance but brand it as a mobile 5870.

You have to throw away everything you know about the desktop model numbers, and treat the mobile part numbers as completely unrelated.

I see it as a cynical attempt by nv and AMD to mislead consumers who see "mobile GTX 470" then use Google to find a chart of desktop video card benchmarks, not realizing those benchmarks do not apply.

Maybe I didn't state clearly enough, but the issue is not laptop GPUs sharing the model names with desktops counterparts. What I'm pointing out is that for example, if I saw a laptop with a 5850m I can not accurately predict the GPU performance unless the vendor states the clocks and memory type used or there is a review. But the catch is if vendors crippled a GPU with slow clocks/RAM they will definitely try not to disclose the specs. We never deal with such foul play when we buy desktop cards (did you ever worry about the 5770 you buy that could actually be ~40% slower than others?), at least for midrange and higher.

 

[DaveSimmons]

^ that too, but I'll still also rant about the part numbering 🙂

 

[CurseTheSky]

Unfortunately, it's marketing. It looks a lot better to simply say "HD 5650 graphics!" in the product headline than "HD 5650 graphics @ 450MHz..." or "HD 5630 graphics!" when your competitors says HD 5670 instead. Your average consumer is fairly dumb when it comes to buying computers, so they just go by what the stickers say. If the model number on the graphics card is exactly the same, they MUST perform the same, right? :sneaky:

The Mobility HD 5850 is one of the biggest offenders - the GDDR3 version performs ok, but is nothing to write home about. The GDDR5 version is very capable, on the other hand (roughly equal to a desktop HD 5770 from what I've heard).

Nvidia's product numbers between various generations are a bit confusing too. We all know / expect the GTS 260M to be slower than the GTS 360M, for example, and the GTX 260M to be slower than the GTX 460M. However, which would you expect to be faster, the GTS 360M or the GTX 260M? It may be GTS vs. GTX, but the 200 series is a generation older after all. As it turns out, the GTX 260M is faster (Who knew?) According to the GPU list on notebookcheck, the 9800M GT and 8800M GTX both beat out the GTS 360M, and yet the GTX 285M is beaten by the GTX 460M. That certainly confuses the crap out of me.

 

[SlowSpyder]

While I think AMD and Nvidia certainly do this on purpose, we just have to remember that a mobile GPU may share nothing other than a number with the desktop part with that same number. I also wish that they would make things a little more clear.