frozentundra123456
Lifer
- Aug 11, 2008
- 10,451
- 642
- 126
If you want to do graphically intensive work like gaming, you need a discrete card with either CPU. If you don't either igp is more than adequate. Nobody can predict the future with perfect accuracy, but I seriously doubt for the average user who surfs the web, does E-mail and social messaging, that a hd4000 laptop will be made obsolete any time in the forseable future. I have a work laptop with a sb i5 and a discrete gpu. It has optimus, and I do fairly intensive data base work and use it for analysis and replottong of analytical chromatograms. I have intentionally watched to see if the discrete gpu kicks in, and I have never seen it being utilized. In fact for the type of work I do, CPU performance is critical.
On the desktop igp performance is even less relevant. If you need better performance, just add a discrete card of whatever level you need. Again amds apus are still better than what the average user needs and still can be handily beaten by a low/mid range discrete card if graphics power is needed.
There is still even lower , though , that is trolling , what Charlesk
is actualy doing , the poster R0HT1 just above you summarized
my point wich the troll was perpetualy eluding , talking of 12 months
cycle upgrade even he knows that we re talking of the average
consumer habits...
If that's his argument, he's not making it well, and it's not a good argument anyway.
The amount of time a machine lasts before being considered obsolete is directly proportional to how powerful it is now. There is no way to know what will happen in the future, but if you want headroom you plan that based on what is here today.
If he wants to say he thinks Trinity will be better for a longer period than IB, fine -- it's speculation, but he's entitled to his opinion. But Sandy Bridge laptops being outdated before Llano ones only mattered (if you believed that) when those were the latest offerings.
This also ignores the discrepancy in CPU performance, which affects everything, not just gaming.
Furthermore, anyone who cares enough about graphics to worry about their machine becoming obsolete in the future, is going discrete anyway. That's the real problem with AMD trying to sell chips based solely on a better IGP -- there just aren't that many people who care about the difference between AMD's IGP and Intel's, who would be happy with either. It's also harder to advertise -- the average person doesn't really understand GPUs as much as CPUs.
Last edited:
I agree with most of what you've said but you're wrong on the IGP front, in my family(cousins, second cousins et al) there has not been a substantial desktop upgrade/replacement for the last 5yrs though admittedly most of'em have moved onto laptops & weren't avid gamers either, cause graphics is where the future lies & Intel's push on this front(even for desktops) proves it that CPU advances will crawl from hereon so AMD is in a stronger position going forward !
For years I wanted an AMD fusion-based laptop to replace my existing (then) Intel-based laptop. Folks may remember how much I talked about looking forward to it because of how much I loathed my Intel laptop with its buggy-as-crap IGP (low performing is one thing, low performing and buggy is a whole other level of end-user hell).
But when I finally was forced to buy a new lappy I could not find anything from AMD with an APU-style CPU/iGPU that was remotely price-competitive in the same form factor to an Intel-based laptop equipped with Nvidia discrete mobile graphics (Optimus).
So that is what I have now, a SB based laptop, 17" form-factor, with an Nvidia discreet graphics card. And it works remarkably well, I had no idea Optimus worked so seamlessly until I had one in hand.
The price gap really caught me by surprise, it was about $200 more to get the AMD lappy with comparable gaming performance to the Nvidia mobile stuff.
Now that's been about a year now, so hopefully AMD has found a way to not just make their APU's price-competitive, they need to convince the OEMs to make the final products themselves price-competitive too.
That's the problem, its not AMD at fault for bundling an A10-4600M with Radeon 7670M ! I mean that's totally retarded & more like they don't wanna sell(not insinuating Intel here
) AMD based stuff D:The hd4000 is basically physically identical between chips. The only variation is the clockspeed (and cache though that is more cpu related). That is why they use the same name. In comparison, amd apus are different physically between models. The high end has 384 shader cores, mid range is 256 and low range is 192.
I don't think you have realized that many of the reviews of the hd4000 were done with a i7-3720m (toms, anandtech) which has a max gpu clock of 1250 mhz. Between different SV models their is really minimal differences except for cache (excluding ultra high end models such as the 3820qm which can't really be found on prebuilt notebooks). This is of course excluding ULV.
3610qm-1100mhz --very popular but refereshed to 3630qm
3612qm-1100mhz--35 watt
3615qm-1200mhz
3630qm-1150mhz--probably the most popular
3632qm-1150mhz--35 watt
3635qm-1250mhz
i5-3360m--1200 mhz
i5-3320m--1200mhz
i5-3210m--1100mhz
i3-2120m-1100mhz
i3-2110m--1100mhz
36xx series have 6mb cache, i5 and i3 have 3/2. You can clearly see that the low end i7s (which are really the only ones outside of boutique notebooks) have very comparable speeds to the lower end i5 and i3 models. Cache is really the only difference and performance is quite similar (note also that in a notebook performance can vary a bit because of issues between turbo-boost and thermals).
Phynaz
Lifer
- Mar 13, 2006
- 10,140
- 819
- 126
I equate the "HD4000" label as being like a brand, same as labeling a CPU "Core i7".
No two Core i7's are going to perform the same (unless they are the same SKU), and I wouldn't expect a HD4000 from one SKU to be the same as the HD4000 from another SKU.
I do agree with AtenRa that AMD's SKU labels are more informative than Intel's.
I don't like that I can't figure out how the HD4000 iGPU is configured to perform just from reading the SKU label.
Intel has 4-digits there (HD4XYZ), the first being reserved to indicate the generation, but those three zeroes could be incremented to give some indication of the iGPU configuration. An HD4050 vs HD4100 vs HD4350, etc.
That Intel didn't do that, when they so easily could have, suggests intentional marketing obfuscation is afoot, and I don't care much for intentional marketing obfuscation.
Interesting, I wonder if any of the buyers out there looking for HD4000 based parts realize that Intel IGP's on lower SKU's perform worse & throttle more aggressively than the top end parts that most reviewers have tested !
Last edited:
Proves nothing of the sort. A longer upgrade cycle doesn't mean AMD is in a stronger position. If anything, people changing platforms less frequently means they can get a boost by putting in a new video card, making the IGP irrelevant.
I'm also betting your cousins also all use dGPUs.
Cheaper parts running at lower clock speeds perform worse than more expensive ones running at higher clock speeds? Yeah, I think buyers can figure that out.
And BTW, while AMD certainly does a better job of differentiating their IGPs, they too sometimes use the same model number with different speeds.
Its kinda easy.
in general i7>=i5>=i3
between classes
take the i7-3610qm
as the numbers increase, in general then gpu clockspeed increases (with a few exceptions).
for i5
Look at second digit from the right.
3360>=3320>=3310.
SV parts have enough headroom not to throttle (assuming no thermal issues--which is the manufacturer's fault). Its ULV that does have massive problems with throttling.
Nope, that's not the point, what I'm saying is that these days "its good enough" is more than adequate for 80~90% of the PC industry & more people are upgrading out of necessity cause the costs for incremental performance upgrades are becoming prohibitively expensive !
What this means is that we're moving towards the mobile industry model of "chip" as a commodity & IGP is playing the most important role in that single component. AMD with its historic lead in this dept, especially after acquiring ATI, is in a better position to leverage this sea change however the traditional x86 market is gonna shrink further whether one likes it or not !
The number of users(including corporates) actively upgrading their systems is barely 10~20% of the total install base of x86, the number of users having dGPU is less than half of that & this number is further going down because people are replacing their desktops with notebooks, see the various sub forums for instance, therefore the role of IGP is not diminishing & in fact its actually the reverse !
Not many of them would know how an HD4000 on an i7 SKU would perform wrt an i3 based HD4000 part, now would they ?
Last edited:
AtenRa
Lifer
- Feb 2, 2009
- 14,003
- 3,362
- 136
All of the above Intel CPUs state they have HD4000. That means that the iGPU in everyone of them must have exactly the same performance. We both know very well that this is not the case, 45W HD4000 doesnt perform the same as 17W HD4000.
Have a look at AMDs iGPU naming
A10-5750M 35W 384 Radeon Cores at 720MHz /533MHz - GPU name = HD 8650G
A10-4600M 35W 384 Radeon Cores at 685MHz/496MHz GPU name = HD 7660G
A10-4655M 25W 384 Radeon Cores at 496MHz/360MHz GPU name = HD 7620G
A8-4555M 19W 384 Radeon Cores at 424MHz/320MHz GPU name = HD 7600G
All of the above AMD iGPUs have the same number of Radeon Cores, would you like AMD to name them all as HD 8000 and be done with it ??? Would you or would you not have a problem with that ?? I would and I believe everyone should have a problem because it would be misleading to the uneducated consumer.
Intels iGPU naming is misleading and they should change it.
If the CPU is "good enough" then the GPU is "good enough" as well.
CPU capabilities are easier to advertise, and have more of an impact on most uses of a machine other than gaming, and for gaming, a discrete card is better anyway.
I see no reason to believe that the IGP is "the most important". Again, other than gamers, most consumers don't even know what video hardware they have.
Because they're not stupid, and they understand that more expensive things perform better than less expensive things? Of course, that's to the extent they care at all. You can't argue that people are buying things based on "good enough" and then also try to claim that the differences in HD4000 performance between chips is a major issue. It's not.
What's the name of the IGP in the A10-5700? What speed does it run at?
What's the name of the IGP in the A10-5800K? What speed does it run at?
How about the A4-3300 and A4-3400? Same name, 35% difference in speed.
Last edited:
AtenRa
Lifer
- Feb 2, 2009
- 14,003
- 3,362
- 136
We are talking about mobile, but you are right both AMD and Intel Desktop iGPUs need better naming.
It's less important to confuse desktop buyers than laptop buyers?
Oh, and there are also examples from Llano, Zacate, Ontario, etc.
That makes absolutely no sense, especially on AT where GPU is the one that matters more !
Again a fallacy, GPU is becoming more & more important & the single most used app(browser) out there is GPU bound these days.
IGP is not the most important thing, its graphics & with things going mobile dGPU will become less relevant ergo the rise of SoC hence IGP.
You're again deflecting the part about Intel misleading(perhaps unintentionally) buyers that the IGP(HD4000) on an Intel i3 is not that much worse than the one on a mobile i7 chip otherwise they ought to have named them differently.
AtenRa
Lifer
- Feb 2, 2009
- 14,003
- 3,362
- 136
No, i have said they both need to fix their iGPU naming both in Laptops and Desktops.
But there is more of a problem with Intel iGPUs naming in laptops.
You're just repeating a claim with nothing to back it up, and we weren't talking about just AT.
Prove it.
Oh, forget it. People like you are never going to discuss this subject honestly -- you're more interested in pushing an agenda.
The author is a self-important dilettante.
Insults are not allowed in technical forums
Markfw900
Anandtech Moderator
Insults are not allowed in technical forums
Markfw900
Anandtech Moderator
Last edited by a moderator:
What agenda are you talking about ?
More people watch bluray movies & youtube videos, play flash based games, browse the web on their notebooks as compared to folks playing crysis 3 or using AutoCAD !
Now if you're saying that the IGP doesn't matter in any of these cases then that's the end of debate right there !
Also if you're not aware of GPU acceleration on apps such as media players, browsers then its better to inform yourself than calling someone a shill & I don't need to prove those numbers to you because if you're a mod(well informed one) here on AT you should be well aware what the rest of the world does & not limit your myopic view to the western world !
First of all we are talking about laptops here, not desktops and thermals are a big deal. Generally a 10% spread (from min to max) is not uncommon for laptops using the same cpu in cpu benchmarks. Other factors such as RAM are also going to impact the igp. So between two laptops with the same igp there may still be a substantial difference.
Second of all, I have mentioned before that you can pretty well extrapolate gpu performance from the model numbers.
Haswell has a number of different model numbers for the igp.
Furthermore the spread between amd igps is much larger. Between an i7 hd4000 (for example the 3630qm) and a ULV i5 there may be a 40% difference. There is a much larger difference between an a10 SV and an a4 ULV.
I agree that the naming system is confusing but AMD isn't really helping consumers either with the billion names they have for their igps. (What I mean is that most consumers are going to look at the laptop and see a8 or a10 and assume that the a10 has a better igp than the a8, rarely will they look at the confusing igp model number and make their decision there). For the same reason, I don't think many consumers assume that their ultrabook is going to have the same igp performance as a larger i7 notebook.
AtenRa
Lifer
- Feb 2, 2009
- 14,003
- 3,362
- 136
Intel Haswell will double the iGPU die size over IvyBridge at the same node.
I believe that proves it
Intel knows that a lot of applications can execute way faster within the GPU than the CPU, they constantly upgrading their iGPUs in order to stay competitive against AMDs APUs. In the mainstream market, you will see the APU(iGPU) wars from now on.
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 23K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
-
Facebook
Twitter