Will there be a 'new' manufacturers race; to 'Fusion'?

[Amiga500]

Again, biased opinions based on no actual facts.

Are you seriously trying to argue that any Intel based IGP can go toe-to-toe with a 780G for playing games?

 

[Wisecracker]

How does what you posted prove that GM45 is lacking high definition content playback capability? How does that prove X4500HD is lacking in performance?

I posted earlier, with real benchmarks, from notebookcheck:
http://www.notebookcheck.net/Intel-Graphics-Media-Accelerator-4500MHD-GMA-X4500MHD.9883.0.html

Not only X4500HD outperformed the entry level discrete graphic cards from last year, its almost on par with Nvidia's last generation mid-level discrete graphic card (GeForce Go 7400). ~~snip~~

X45 = Underwhelming and Overhyped.

""Targeting"" 3X Performance = Intel ""Marketing""

Nice try in rolling out 2 year-old tech for your comparison. What's interesting is that in your link (for one example) the x45 in 3dmark06 uses a resolution of 1280x800 instead of the standard testing resolution of 1280x1024 to gain an estimated 11% increase overall.

On the other hand, ATi's IGP doesn't seem to improve that much at all, if 3DMark06 is of any indication. The difference between 690G (HD2400 core) and 780G (HD3200) is a measly 300 points.

Weak. Lame. And. Feable. (Just like the GMA G/X 45)

 

[sarwar_r87]

never in my life i though i would see people trying to defend a IGP made by intel..........i am amused.....plz do carry on....

NOTE: srsly, i dont think we need graps to prove intel's dominance in IGP, jus ask anyone one the streets. in this case words are more than enof

 

[Reynod]

This is why I can't take any of the intel fanboys seriously.

Their current CPU's are great single socket chips .... that however says nothing about their IGP solutions .... which are just utter utter crap.

Crap drivers too.

They are not even worthy to be compared to Matrox let alone Nvidia and ATI.

Why are we discussing "Office Machines" ???

yomamafor1 ... you should just give up ... the other pets won't support you on this one ... the benchies can't be argued with I am afraid.

 

[archibael]

X45 = Underwhelming and Overhyped.

http://www.dvhardware.net/news/intel_g45_slide.jpg

http://www.notebookjournal.de/storage/show/image/image482aa5a547d8c_text

""Targeting"" 3X Performance = Intel ""Marketing""

The marketing "3X" is compared to G33, which is the GMA3100-- has no HW support for T&L. The X3100 is GM965-- mobile G965, which has HW T&L. Unfortunately, I don't see G33 in that comparison. Not saying it's that much better (I actually have no idea), but those graphs do not support the conclusion you are drawing.

 

[BaronMatrix]

How does what you posted prove that GM45 is lacking high definition content playback capability? How does that prove X4500HD is lacking in performance?

I posted earlier, with real benchmarks, from notebookcheck:
http://www.notebookcheck.net/Intel-Graphics-Media-Accelerator-4500MHD-GMA-X4500MHD.9883.0.html

Not only X4500HD outperformed the entry level discrete graphic cards from last year, its almost on par with Nvidia's last generation mid-level discrete graphic card (GeForce Go 7400).

Is it strong by any means? Absolutely not. Is it competitive against AMD's 780G? Not at all. What X4500HD signify, however, is that Intel is already making significant improvement on its IGP technology.

On the other hand, ATi's IGP doesn't seem to improve that much at all, if 3DMark06 is of any indication. The difference between 690G (HD2400 core) and 780G (HD3200) is a measly 300 points.

Here is all the proof you need that X4500 is more smoke and mirrors from Intel.

http://blip.tv/file/1129114

 

[NMDante]

Isn't it strange that Aero is turned off on the AMD laptop, and sidebar not running, while the Intel has Aero turned on, and the sidebar running.

Yeah, that's smoke and mirrors.

Thanks for showing the deception behind a supposed "equal" test.

Oh, and you can tell Aero is turned off, since the Windows taskbar in the AMD machine is not transparent, but solid, while the Intel taskbar has the Aero transparency going.

Why not show the Windows display settings before running this "test"? Show that both laptops are running the same resolution, same Aero/non-Aero settings, with the same sidebar/no sidebar running?

Good try, though.

 

[BaronMatrix]

Isn't it strange that Aero is turned off on the AMD laptop, and sidebar not running, while the Intel has Aero turned on, and the sidebar running.

Yeah, that's smoke and mirrors.

Thanks for showing the deception behind a supposed "equal" test.

Oh, and you can tell Aero is turned off, since the Windows taskbar in the AMD machine is not transparent, but solid, while the Intel taskbar has the Aero transparency going.

Why not show the Windows display settings before running this "test"? Show that both laptops are running the same resolution, same Aero/non-Aero settings, with the same sidebar/no sidebar running?

Good try, though.

Well, I guess you can get mad at them. Let's assume that Aero DOESN'T REALLY USE MUCH CPU. Perhaps I should link you to a test of Half Life 2 where Puma scores 50fps and Centrino scores 14fps (@ 1024). Just search Puma on YouTube. There is also another one that shows WOW with NO DX9 textures or transparency. Face it, Intel graphics are like a retarded kid with crayons and X4500 has done nothing to change it.

Oh alright. I'm sure you'll find something wrong there too. Knock yourself out.


http://www.youtube.com/watch?v=2VcJojNMtmE Gaming/HD on desktop

http://www.youtube.com/watch?v=n60bafkGjew Gaming on laptop

These are all 780G vs G45

 

[yomamafor1]

I give up. I guess some just can't throw away the prejudice, and focus on the improvments.

 

[yomamafor1]

Well, I guess you can get mad at them. Let's assume that Aero DOESN'T REALLY USE MUCH CPU. Perhaps I should link you to a test of Half Life 2 where Puma scores 50fps and Centrino scores 14fps (@ 1024). Just search Puma on YouTube. There is also another one that shows WOW with NO DX9 textures or transparency. Face it, Intel graphics are like a retarded kid with crayons and X4500 has done nothing to change it.

Oh alright. I'm sure you'll find something wrong there too. Knock yourself out.


http://www.youtube.com/watch?v=2VcJojNMtmE Gaming/HD on desktop

http://www.youtube.com/watch?v=n60bafkGjew Gaming on laptop

These are all 780G vs G45

Can't you provide something more solid than... pathetic home-made youtube videos?

Yes, 780G is much superior than G45. That's a fact since the beginning.

Its ironic, that AMD has the better interconnection, and arguably platform, yet has much inferior CPU. Intel on the other hand, has less powerful interconnection, and arguably platform, but have much stronger CPU.

The point is, Intel is already making significant improvement on its IGP and platform, but is AMD making significant improvement on its CPU?

 

[BaronMatrix]

I give up. I guess some just can't throw away the prejudice, and focus on the improvments.

You mean like how Phenom is faster clock for clock than K8 (on the heels of Penryn) and 780G is the greatest IGP second only to 790GX and SB750 is the best OC chipset right now?

 

[yomamafor1]

SB750? The best OC chipset? You mean the best, also the latest, AMD Southbridge?

:lol:

 

[BaronMatrix]

The point is, Intel is already making significant improvement on its IGP and platform, but is AMD making significant improvement on its CPU?

You mean now it's liek two retarded kids with crayons? AMD has improved from K8 over 25% considering Deneb. Perhaps more by the time it ships.

 

[yomamafor1]

25%? More like 18% clock for clock. It also looks like Deneb's clockspeed will be around Phenom as well.

But of course, we should wait for the real benchmarks to arrive, instead of opiniating and speculating.

 

[Slobogob]

You mean like how Phenom is faster clock for clock than K8 (on the heels of Penryn) and 780G is the greatest IGP second only to 790GX and SB750 is the best OC chipset right now?

There is no such thing as a 790GX right now. I would've bought it, but it is not for sale. It's AMdelayed.

 

[BaronMatrix]

25%? More like 18% clock for clock. It also looks like Deneb's clockspeed will be around Phenom as well.

But of course, we should wait for the real benchmarks to arrive, instead of opiniating and speculating.

But you won't even accept that improvement so what difference does it make? Super Pi IS NOT indicative of the average desktop use. You may as well use 3D Mark.

 

[BaronMatrix]

There is no such thing as a 790GX right now. I would've bought it, but it is not for sale. It's AMdelayed.

It was due to ship in the last week of July. ASUS press-released one which means it's on the way to Newegg. I actually thought it would show up by now. I'd say early next week. They should all make it to retail.

 

[yomamafor1]

But you won't even accept that improvement so what difference does it make? Super Pi IS NOT indicative of the average desktop use. You may as well use 3D Mark.

Baron, its called "correcting your overinflated figures", not "not accepting improvements". If I didn't accept any improvements, would I be acknowledging there's improvement in the first place?

Super Pi is not indicative of average desktop use, but its a preliminary performance comparison between K8 and K10, no? That's all we've had at the moment in regards to Deneb's performance.

Again, as I said before, we should wait until the real performance benchmarks, not going out and spewing FUDs like "25% faster".

 

[yomamafor1]

It was due to ship in the last week of July. ASUS press-released one which means it's on the way to Newegg. I actually thought it would show up by now. I'd say early next week. They should all make it to retail.

790GX will show up in mid-Aug.

 

[jimmysmitty]

25% with a die shrink. I will only believe it when I see it. At most I would expect Deneb to give 10-15% with the die shrink and thats IF it goes well.

Problems are possible. But 25%? Thats a wish and a dream.

 

[Amiga500]

25% with a die shrink. I will only believe it when I see it. At most I would expect Deneb to give 10-15% with the die shrink and thats IF it goes well.

Problems are possible. But 25%? Thats a wish and a dream.

I am assuming that the move to K10.5 is more akin to the RV600 to RV670 move than normal die shrinks - there will be detail march changes. Intel made some quite substantial changes to Conroe when moving to Penryn - look at the change in performance for FPU divisions:

The only weakness remaining in the Core x87 architecture is the FP divider. Notice how even a relatively low percentage of divisions (the 4th number in the mix) kills the performance of our 65nm Xeon. The Opteron 22xx and 23xx are 70% faster (sometimes more) when it comes to double precision FP divisions. However, the new Xeon 54xx closes this gap completely thanks to lowering the latency of a 64-bit FDIV from 32 cycles (Xeon 53xx) to 20 cycles (Xeon 54xx, Opteron 23xx). The Xeon 54xx is only 1% to 5% slower in the scenarios where quite a few divisions happen.

They (AMD) have indicated notable IPC improvements, and that will not be the result of a die shrink.



I do agree with your 15% estimation for K10.5, 25% would be a bit much (although not impossible if K10 has some specific problems holding the whole thing back)

 

[jimmysmitty]

^True but I would have expected AMD to get those problems out with the B3 stepping. That is unless it is the 65nm stepping and 45nm allows the arch to shine like it should have.

In that case Intel would have been the smarter one in waiting to go naitive quad at 45nm instead of 65nm.

Oh and I always thought AMD was not as cache sensative as Intel.

But only time will tell.

 

[BaronMatrix]

Baron, its called "correcting your overinflated figures", not "not accepting improvements". If I didn't accept any improvements, would I be acknowledging there's improvement in the first place?

Super Pi is not indicative of average desktop use, but its a preliminary performance comparison between K8 and K10, no? That's all we've had at the moment in regards to Deneb's performance.

Again, as I said before, we should wait until the real performance benchmarks, not going out and spewing FUDs like "25% faster".

First I said 25% over K8 not K10. You said they haven't improved. That's FUD.

 

[spud]

Here is all the proof you need that X4500 is more smoke and mirrors from Intel.

http://blip.tv/file/1129114

There is a color scale difference, the video on the left is of lesser quality as well.

Word, Playa.

 

[yomamafor1]

First I said 25% over K8 not K10. You said they haven't improved. That's FUD.

In my exact words:

25%? More like 18% clock for clock. It also looks like Deneb's clockspeed will be around Phenom as well.

Did I say they did not improve? No, I said you overinflated the figures by 7%.

I didn't know "twisting words" is also part of the screenplay writing.