Engadget: AMD to demo "Zacate" next week.

[extra]

You do not know me at all . My daughter will sell to anyone . NOT me and none could stop it, This wasn't racist remark . I sold only expensive gamers.

expensive organically grown gamers always taste the best, if you know what I mean.

Anyway, there's a big thread over at semi accurate that has more info on the demo.

 

[Vesku]

How is that a problem? That SB was a desktop part, Apples and Oranges and all that. My main take away is that AMD did a decent job with Bobcat. I'm hoping that some OEM jumps the gun on Zacate, I happen to be in the market for a low budget notebook in November/December time frame.

I do agree with the sentiment that Anand should have run some CPU limited tasks, though. My guess is AMD is wary of that because they don't want to show their Ontario vs Atom hand yet.

 

[busydude]

How is that a problem? That SB was a desktop part, Apples and Oranges and all that. My main take away is that AMD did a decent job with Bobcat. I'm hoping that some OEM jumps the gun on Zacate, I happen to be in the market for a low budget notebook in November/December time frame.

I guess Nemesis's argument is that, since all mobile SKU's have 12 EU GPU parts, they should beat Zacate's GPU in terms of performance.

 

[Idontcare]

Check out the Zacate update, http://www.anandtech.com/show/3933/amds-zacate-apu-performance-update , AMD gave Anand full control of their demo platform. I think that in itself is worth a gold star. Unlike Nemesis 1, I don't see Intel deploying a SB competitor any time soon in that same value and power bracket. Intel has always protected their margins when it comes to mobile. Although for non-value notebooks it seems like Intel will maintain their comfortable position given Llano delays.

Remains to be seen if Ontario puts a squeeze on Atom prices and volume in the netbook space, but it seems likely. *Edit* Intel is pretty nimble for a 800Lb gorilla, just look how they are trying to leverage atom success in netbooks to squeeze it into embedded and MIDs. Not sure how low Ontario can go considering it's design.

Anand is THE man :thumbsup:

Mad props to AMD too for acknowledging the benchmark discrepancy existed and allowing a 3rd party reviewer to get in there and hash it out to the best of their experted abilities.

I'm actually quite astounded at the chutzpah AMD showed there. Quite a gutsy move, and then they allowed Anand to publish it on top of it all.

Definitely shows the confidence that AMD has in Zacate.

And if little-ole Zacate is this good...holy Llano!

 

[Nemesis 1]

Check out the Zacate update, http://www.anandtech.com/show/3933/amds-zacate-apu-performance-update , AMD gave Anand full control of their demo platform. I think that in itself is worth a gold star. Unlike Nemesis 1, I don't see Intel deploying a SB competitor any time soon in that same value and power bracket. Intel has always protected their margins when it comes to mobile. Although for non-value notebooks it seems like Intel will maintain their comfortable position given Llano delays.

Remains to be seen if Ontario puts a squeeze on Atom prices and volume in the netbook space, but it seems likely. *Edit* Intel is pretty nimble for a 800Lb gorilla, just look how they are trying to leverage atom success in netbooks to squeeze it into embedded and MIDs. Not sure how low Ontario can go considering it's design.

Have you seen the Atom processors due out in 2011 perform . Until we see whats new from intel its a up in the air. How many Atom processors did intel show at IDF how many differant formats .

As I said AMD was simply trying to beat up on the I 5 cpu part . Had nothing to do Graphics at all .

If it was about graphics AMD could have killed 2 birds with 1 stone . And tested against Atom with NV graphics . Than this would have been meaning full. Ya I know its not in a note book . AMD choose to use a notebook rather than a netbook.

Lets take sandy bridge part .how many watts is the IGP using. On the lowest binned part.
25watts 30 watts or less.

Hell it could be intel doesn't even care about the $500 and below parts . But if that were true there would be NO Oak trail parts coming. I not sure whats in oak trail Bob insist its Imagination tech . But the SB part looks good to me.

 

[Nemesis 1]

Anand is THE man :thumbsup:

Mad props to AMD too for acknowledging the benchmark discrepancy existed and allowing a 3rd party reviewer to get in there and hash it out to the best of their experted abilities.

I'm actually quite astounded at the chutzpah AMD showed there. Quite a gutsy move, and then they allowed Anand to publish it on top of it all.

Definitely shows the confidence that AMD has in Zacate.

And if little-ole Zacate is this good...holy Llano!

What else could AMD do IDC . and Anand did report the orginal story . Everyone new they gave a false test. AMD had to let Anand do this with all the egg on both parties faces . Wouldn't you say . But they did a good job of cleaning it up . I gave Anand an A+ for the batman move and AMD I believe was sleeping here like they were on the orginal report. SB flatout eats it alive . So Anand did better than good he did great .

 

[Vesku]

I wouldn't blame AMD on that one Nemesis 1, I setup my sister's Toshiba laptop and it was the same deal graphics wise. Stuck with the OEM drivers from months ago, required some googling and OS know how to remedy.

I think you are on to something with APU being the new AMD64. GPU computing should start to catch on even for consumer software once the install base gets large enough. SB graphics aren't even DX11 let alone OpenCL capable, correct? Although it is still an impressive performance leap for Intel integrated.

 

[Nemesis 1]

I guess Nemesis's argument is that, since all mobile SKU's have 12 EU GPU parts, they should beat Zacate's GPU in terms of performance.

Yes sir and they do . But cost well be hugh differance. Best part only NV fanbois can call the SB igp crap now . None really cares about NV . as they have no x86 parts .

 

[Nemesis 1]

I wouldn't blame AMD on that one Nemesis 1, I setup my sister's Toshiba laptop and it was the same deal graphics wise. Stuck with the OEM drivers from months ago, required some googling and OS know how to remedy.

I think you are on to something with APU being the new AMD64. GPU computing should start to catch on even for consumer software once the install base gets large enough. SB graphics aren't even DX11 let alone OpenCL capable, correct? Although it is still an impressive performance leap for Intel integrated.

Ok . Look as AMD is testing parts . Their testing against the comp also . If threw that testing period . AMD were using broke drivers and didn't no it and thought they had a 10x faster part in 2D . They missed the target. These are highly payed educated people . I don't believe they made a mistake at all . Look at the differance in screen quality between the test systems. Lots of Gulf state properity on the market ya buy some.

 

[Nemesis 1]

I wouldn't blame AMD on that one Nemesis 1, I setup my sister's Toshiba laptop and it was the same deal graphics wise. Stuck with the OEM drivers from months ago, required some googling and OS know how to remedy.

I think you are on to something with APU being the new AMD64. GPU computing should start to catch on even for consumer software once the install base gets large enough. SB graphics aren't even DX11 let alone OpenCL capable, correct? Although it is still an impressive performance leap for Intel integrated.

I not sure about Direct compute we had no luck with it . But ya Intel SB parts are CL1.1 and GL 3 complaint. No tess. But I would like to see this zapata do a game with tess. LOL.

Intel did enough with SB to make Apple happy.

 

[Idontcare]

What else could AMD do IDC . and Anand did report the orginal story . Everyone new they gave a false test. AMD had to let Anand do this with all the egg on both parties faces . Wouldn't you say . But they did a good job of cleaning it up . I gave Anand an A+ for the batman move and AMD I believe was sleeping here like they were on the orginal report. SB flatout eats it alive . So Anand did better than good he did great .

Zacate isn't targeted at segments where its competition will be Sandy Bridge, so what is the relevance of the comparison?

Guess what, GTX480 and the 5970 drink SB's milkshake too...but a comparison wouldn't be relevant so why go there either?

Llano is where its at as far as SB competition in the integrated graphics space.

 

[Chiropteran]

Broken drivers are the reason why I simply won't buy a laptop without an AMD GPU. I know I can use AMD's latest drivers on any laptop I buy (I avoid the brands that disable this functionality), so I don't have to worry about the laptop manufacturer's support being out of date. While the benchmark might not have proven the superiority of the Zacate platform in IE9, it did illustrate a huge problem with most non-AMD laptops, driver support sucks.

 

[Nemesis 1]

Really . Interesting . So how did Anand get AMD to fix the test system was it intels fault . This could never happen to and AMD machine from a PC maker.

 

[Martimus]

Have you seen the Atom processors due out in 2011 perform . Until we see whats new from intel its a up in the air. How many Atom processors did intel show at IDF how many differant formats .

As I said AMD was simply trying to beat up on the I 5 cpu part . Had nothing to do Graphics at all .

If it was about graphics AMD could have killed 2 birds with 1 stone . And tested against Atom with NV graphics . Than this would have been meaning full. Ya I know its not in a note book . AMD choose to use a notebook rather than a netbook.

Lets take sandy bridge part .how many watts is the IGP using. On the lowest binned part.
25watts 30 watts or less.

Hell it could be intel doesn't even care about the $500 and below parts . But if that were true there would be NO Oak trail parts coming. I not sure whats in oak trail Bob insist its Imagination tech . But the SB part looks good to me.

It sounds like you have a major problem with AMD from your posts, but maybe there is something to them. Are you trying to say that there will be a low power Sandybridge competitor to Ontario/Zacate? If so, will I be able to buy a notebook/netbook at around $300-$500 to replace the current Intel laptop I own (that has unacceptable graphics that continuously cause it to crash when I am using it to run my Eve client)? If so, then I would be interested.

If that is not what you are saying, then why are you talking about Sandybridge at all? The market that Ontario/Zacate is in, is for cheap laptops; not high priced laptops. The markets are completely different, so why compare them? AMD is releasing a product for this particular market, and it appears to be better for that market than the Intel competition. This is not surprising really, since Intel never intended for the chip to be used in the "netbook" market, so unlike Ontario it wasn't really designed for it.

 

[cbn]

Check out the Zacate update, http://www.anandtech.com/show/3933/amds-zacate-apu-performance-update , AMD gave Anand full control of their demo platform. I think that in itself is worth a gold star. Unlike Nemesis 1, I don't see Intel deploying a SB competitor any time soon in that same value and power bracket. Intel has always protected their margins when it comes to mobile. Although for non-value notebooks it seems like Intel will maintain their comfortable position given Llano delays.

Yep, It is still quite amazing to me that an 18 watt Processor (built on 40nm nonetheless) is showing equivalent results in IE9 benchmark compared to a 35 watt processor built on 32nm.

 

[Accord99]

Yep, It is still quite amazing to me that an 18 watt Processor (built on 40nm nonetheless) is showing equivalent results in IE9 benchmark compared to a 35 watt processor built on 32nm.

It's AMD's future 40nm integrated GPU showing equivalent browsing performance and better gaming performance than Intel's current 45nm integrated GPU. Given the nature of Intel graphics, this is pretty much expected.

The processors don't really come into play in these tests and the power comparison is misleading seeing as how an i5-520m will probably have 3 times more CPU throughput than an 1.6GHz Zacate.

 

[Nemesis 1]

Zacate isn't targeted at segments where its competition will be Sandy Bridge, so what is the relevance of the comparison?

Guess what, GTX480 and the 5970 drink SB's milkshake too...but a comparison wouldn't be relevant so why go there either?

Llano is where its at as far as SB competition in the integrated graphics space.

A little double standard here IDF , Its not meant to go against SB . I already said that yesterday , Clearly stated thus my rant about the I5 part being tested . I5 wont be in 2011 notebooks either. So its the 2011 parts that matter. AMD has a small window of oppertunity here . What part 22nm is Intel going to release 2nd half of 2011. If JF can use the argument BD in 2011 . So I can say what intel said 2nd half of 2011. That means july right .

 

[Nemesis 1]

Yep, It is still quite amazing to me that an 18 watt Processor (built on 40nm nonetheless) is showing equivalent results in IE9 benchmark compared to a 35 watt processor built on 32nm.

Ya its amazing that an intel IGP in IE9 is = to ATI 5970 a 35 watt part up against a 225 watt GPU.

 

[cbn]

It's AMD's future 40nm integrated GPU showing equivalent browsing performance and better gaming performance than Intel's current 45nm integrated GPU. Given the nature of Intel graphics, this is pretty much expected.

The processors don't really come into play in these tests and the power comparison is misleading seeing as how an i5-520m will probably have 3 times more CPU throughput than an 1.6GHz Zacate.

Thanks for clearing that up. If Gaming performance is greater than browser performance does that mean Browser GPU optimization still has a way to go? Or is this more likely due to the fact Web pages are 2D and games are 3D?

P.S. Hopefully we can get some more testing on actual web page loading time/web surfing performance. I think it would be interesting to see how much "GPU" factors into the equation compared to CPU?

 

[Accord99]

Thanks for clearing that up. If Gaming performance is greater than browser performance does that mean Browser GPU optimization still has a way to go? Or is this more likely due to the fact Web pages are 2D and games are 3D?

I think it's could be Direct2D is already well accelerated by everything; I tried it on a X4500 from the Intel GM45 chipset which score around 1650 so even lousier Intel graphics doesn't have much problems with it.

 

[Vesku]

What would you have them demo Zacate with Nemesis 1? Pentium mobile dual core?

 

[Idontcare]

A little double standard here IDF , Its not meant to go against SB . I already said that yesterday , Clearly stated thus my rant about the I5 part being tested . I5 wont be in 2011 notebooks either. So its the 2011 parts that matter. AMD has a small window of oppertunity here . What part 22nm is Intel going to release 2nd half of 2011. If JF can use the argument BD in 2011 . So I can say what intel said 2nd half of 2011. That means july right .

I'm with ya, I'm with ya. Yer sayen that if AMD didn't want Zacate compared to Sandy then they wouldn't be out there comparing it to i5 themselves.

Fair enough. I see your beef, and trajectory with this line of thinking (AMD made the comparison, not Intel and not Nemesis).

 

[Vesku]

Well, I guess they could have gone i3 but if they'd gone with anything lower they'd probably get criticized for using an older intel integrated. Doubt Intel will lend them a low power SB mobile platform to compare with. ;p

 

[cbn]

I think it's could be Direct2D is already well accelerated by everything; I tried it on a X4500 from the Intel GM45 chipset which score around 1650 so even lousier Intel graphics doesn't have much problems with it.

It looks like I have been guilty of not reading my own thread. As early as post #52 I am seeing comments and criticisms that IE9 benchmark is already saturated by a wide variety of GPUs.

So now I am wondering how much contribution the GPU really has with respect to web page loading times? Is it great or small? (I'm guessing small unless the processor happens to be low powered ARM)

 

[386DX]

Thanks for clearing that up. If Gaming performance is greater than browser performance does that mean Browser GPU optimization still has a way to go? Or is this more likely due to the fact Web pages are 2D and games are 3D?

P.S. Hopefully we can get some more testing on actual web page loading time/web surfing performance. I think it would be interesting to see how much "GPU" factors into the equation compared to CPU?

It means that the IE9 browser test is pretty much useless as a comparison for graphics performance because like others have said with "proper" drivers an old Intel IGP = Zacate = Radeon 5970. And for everyone defending AMD by saying its an "honest" mistake that they weren't using the latest Intel drivers for there tests you're pretty naive.

AMD is a tech company full of technically minded people. They have an entire division to just write drivers for there video cards so they know how much drivers can improve performance. The fact they listed the components for both system on there demo yet they failed to disclose the video driver version when doing a "graphics" comparison is shady. Look at when Anandtech or any other tech site does reviews of video cards, they always disclose what driver version were used in the test, and they pretty much always use the latest driver and never the OEM drivers that came with the video card. So for AMD to use the laptop manufacturer didn't have an updated video driver as an excuse is pretty weak.

Onto Zacate itself. There's no doubt that Zacate is gonna offer a significant graphics improvement over Atom and its IGP. Whats more troubling to me is when you demo your product you're obviously gonna show it off in the "best" possible scenario. If the "shady" IE9 demo, and COH and the few others Zacate was running was the best they got then even though the graphics have improved its not great. Sure its faster then the i5 at Batman, but 16 fps or 12 fps its still unplayable, its why they didn't demo it with that game or other more "popular" games. The point is, whether its 50% faster then the IGP in the i5 or not is really irrelevant if that 50% faster still doesn't let you run the majority of games at a playable frame rate.

Which brings me to the last point. There was virtually no mention of the CPU performance. All the focus has been on the GPU side. The only numbers that have been seen was that a bobcat core would be about as fast as a Core 2, I believe the benchmark listed the Core 2 at 1.6 GHz and an Atom at 1.66 GHz, however the bobcat frequency was not disclosed. This is troubling to me because in the past AMD has done this. Claiming so and so is "as fast" or "about the same speed as" and it always turned out there claims were based not on a clock for clock comparison but a higher clocked AMD part to a lower clocked Intel part. This leads me to believe the CPU side of Zacate is gonna be alot weaker then what people are expecting.