Llano Demo

[ElFenix]

Just tell me here and now what speed llano was running at . What speed was intels IGP at was , turbo on or off What speed was llano graphics running at . The video may be legit and all but lets see how Intel actually runs on that game , That was a flash game was it not or did i miss something . As far as I know intel wasn't into gaming with IGP until the first HD graphics the first of which was a step forward HD3000 was a leap forward and IB which follows closely behind the llano release will be another leap , Even tho AT has said it will only be 16 shaders Which may be true or not , But I wagers its clockspeed will be very close to 2ghz.

If you think intel won't show results for a real working IB your sadly mistaken . When llano is released intel will show performance at fall IDF of IB . llano is a 3rd qt release is it not?

holy crap dude just give it up

 

[itsmydamnation]

Name me one post or topic were BD/llano have been shown in real reviews

what does that have to do with anything, no one has talked about performance figures, but just look at the architecture (BD) itself it looks the goods. Watch the hotchips presentation of the BD core, the speaker emphasizes its single thread performance several times.

. Ya zacata looks alright . But not in this range. The 5 watt i haven't seen any performance numbers .

take the C-50 numbers

you best check the title of this topic .

i was doing nothing but responding to your post.

Its a llano vs SB AMD marketing video . Its a hype topic with nothing legit about it.

not really, it shows that SB ia lackluster in GPU performance compared to Llano, this can be shown by running multiple GPU based apps at the same time or using higher settings, kind of not fair for a side by side demo to be using non comparable settings, right?

http://www.semiaccurate.com/forums/showpost.php?p=100768&postcount=71

Correct, but my posts have always been my own opinions. Well informed opinions

Llano is to Ridgeback as Penryn is to Conroe. Except, I would argue, a little more than Penryn, but not enough to Nehalem.

if we believe redPrest who worked at amd( on the technical side not marketing) then llano is a little bit more of a performance jump then conroe to wolfdale you can take the exsisting AMD Athlon II(maybe phenmon II, hard to tell, ridgeback has L3) 45nm processors and add 10% performance.

the question then becomes what do mainstream users need more of, CPU performance or GPU performance. Now consider that question with the CPU performance of SB to an athlon II + 10% on the CPU side. consider SB GPU performance even compared to zactate let alone something with 5, 6 times the performance.

to me that's a very simple question to answer.

You act as if the cheap mobill market is going to be a big hit . If zacate was all that even @ 5 watts . AMD should not have gotten rid of DIRK. He went for a reason . Zacate at 5 watts should have been a great lifeline. LLano according to this video kills SB. and you guys and many others have BD for the WIN . SO why the hell is DIRK and friends gone .

i dont have BD for the win, i have never commented on performance, i've only ever tried to apply logical analysis with what we know and that is limited by my technical knowledge. why Dirk is gone, could be lots of reasons, unless Dirk or the board comment on exactly why we can only speculate.

 

[Nemesis 1]

.. and you post a video showing E-350 Vs P6000 and blast the marketing team. I don't understand why you get so riled up..

NO no I not riled up at all. LLano is going to make discrete graphics disapear. NV needs to get out of desktop and intel needs to give up because llano is a game changer . Until you go discrett at which point SB sucks big time comparred to llano with discrete . I didn't know llano was on a BD core.

So NV will destroy Intel in the tablet and small device markets and Zacata llano and BD will make intel stuff useless . Because I will buy an AMD BD if it creams SB in the $200 price range. I like intel but not that much . Is whats amusing to me is that I get the feeling you guys think amd will keep the same price points if its better than SB . If thats true count me in as an AMD user.

 

[Mopetar]

Just tell me here and now what speed llano was running at . What speed was intels IGP at was , turbo on or off What speed was llano graphics running at . The video may be legit and all but lets see how Intel actually runs on that game , That was a flash game was it not or did i miss something . As far as I know intel wasn't into gaming with IGP until the first HD graphics the first of which was a step forward HD3000 was a leap forward and IB which follows closely behind the llano release will be another leap , Even tho AT has said it will only be 16 shaders Which may be true or not , But I wagers its clockspeed will be very close to 2ghz.

If you think intel won't show results for a real working IB your sadly mistaken . When llano is released intel will show performance at fall IDF of IB . llano is a 3rd qt release is it not?

A Flash game? It says what game it is in the video and it's most certainly not a Flash game.

Also, what does it matter what the Llano chip is clocked at as we've seen time and time again that clock speed doesn't necessarily indicate better performance. They showed how much power each chip was drawing, so even if the Llano part were running at something ridiculous like 5 GHz it's still doing it while drawing less power.

Llano is a Q2 release (A few rumor sites have said April.) according to available information and the fact that AMD has shown off a video of it most likely indicates that it's fairly close to release.

Please at least stop and think before rambling on in such an incoherent fashion.

 

[Nemesis 1]

Quote:
you best check the title of this topic .

i was doing nothing but responding to your post.

I wasn't referring to the O/T I was referring to the topic itself and the video and how test was conducted. What if intels IGP turbo was turned off . You think apples to apples is llano graphics and intel graphics both at 850mgz is Fair . I call it BS. I also know that low level cache speed is important to intels IGP performance . Low level cache runs at processor speed and how you slice the cache we have none of that information . the Z67 chip is going to be fun I will love playing with IGP cache size (slice) . I will love O/Clocking the hell out of it . and it will be fine for an internet machine . But for a gamer don't make ne laugh . I ain't that stupid. I will likely be switching to NV graphics as ATI is no more. And I really don't like NV at all but not crazy about AMD either so I will just buy best performing card single gpu .

 

[Nemesis 1]

A Flash game? It says what game it is in the video and it's most certainly not a Flash game.

Also, what does it matter what the Llano chip is clocked at as we've seen time and time again that clock speed doesn't necessarily indicate better performance. They showed how much power each chip was drawing, so even if the Llano part were running at something ridiculous like 5 GHz it's still doing it while drawing less power.

Llano is a Q2 release (A few rumor sites have said April.) according to available information and the fact that AMD has shown off a video of it most likely indicates that it's fairly close to release.

Please at least stop and think before rambling on in such an incoherent fashion.

Really thats news to me thats great . So llano and BD arrive at same time .

I guess it was the zacate video was flash game

 

[Nemesis 1]

holy crap dude just give it up

Was that video not llano vs. SB . Was not there alot of Oh my thats impressive' I seen nothing impressive from the SB setup at all. Which leads me to question the setup . Is it not a fair thing to question. Its like when AT said we won't see B3 chipset till april on M/Bs I questioned that also . There here now according to another topic . I don't care that supply chain has to build up again point is ya can get the M/Bs now . When we stop questioning what were told or lead to believe than friend it is time to give it up .

 

[SickBeast]

I'm sceptical about this demo. SB obliterates the "stars" CPU core, and its integrated graphics are pretty good as well. AMD has definitely rigged this test.

AMD needs to come out with their Bulldozer core with a good integrated GPU.

 

[Nemesis 1]

If you overclock hte IGP in a laptop, you limit CPU performance further.

The only thing that will save Intel in the laptop space is the use of Optimus and the AMD equivalent. If running a dedicated GPU was as power expensive as it was two years ago, Intel wouldn't have a chance in anything but the budget low end and desktop replacements. The entire middle woudl go to AMD.

I think Apple should have not updated the 13" MBP to sandy bridge and just waited for Llano and announced the new Llano MBP along with the SB MB Airs.

Ya I can't wait to see AMDs version of thunderbolt that Apple is going to be using . I bet AMD spent billions in its development.

 

[itsmydamnation]

I'm sceptical about this demo. SB obliterates the "stars" CPU core, and its integrated graphics are pretty good as well. AMD has definitely rigged this test.

AMD needs to come out with their Bulldozer core with a good integrated GPU.

So SB can obliterate STARS but a 400+SP (ie more then R600 had) based GPU can't obliterate SB GPU :\. SB's DX10 support is token at best, open GL doesn't even work. bobcat based APU's reset the bar for GPU performance in that market, llano is going to do the same in the mainstream market, but it comes down to that question again.

what matters more for the laptop and consumer mainstream space, SB CPU perf over llano or llano GPU perf over SB. I really honestly think its the later.


edit:

Ya I can't wait to see AMDs version of thunderbolt that Apple is going to be using . I bet AMD spent billions in its development.

keep shoveling that dirt son, so how much money do you expect people to actually spend on disk to actually generate that amount of IO?

ps. i design high end enterprise networks for a living, including fully converged (data,san,voip) networks/datacenters. I have a very good idea how much disk that can generate that kind of IO costs both from low user count and high user count. one is really expensive and one is outrageous.

for SB based computer to reach llano GPU performance means discrete video card, that pushes up both power draw and cost.

you can really tell im currently in the documentation stage of this project, ie easily distracted by forums

 

[drizek]

I think GPU performance will win easily IF AMD can actually make use of it for encoding and similar applications.

In gaming, no game is going to be cpu limited in Llano, so stars doesn't matter there. For encoding and decoding, GPU acceleration can save the day.

So what else do normal people even use their computers for? Llano should handle virtually every task competently. I would bet it is going to give about the same level of performance as my desktop with a Phenom II 720 and HD 4830 that I was using up until a couple of weeks ago. It literally will run crysis(at medium settings). More importantly, it will run every unreal engine game and every source game at high settings on laptop resolutions. Thats pretty impressive.

 

[Nemesis 1]

I'm sceptical about this demo. SB obliterates the "stars" CPU core, and its integrated graphics are pretty good as well. AMD has definitely rigged this test.

AMD needs to come out with their Bulldozer core with a good integrated GPU.

You know the best part of the video and I am amused by is was the walk threw the guy seemlessly walked threw the llano setup and never lost a step going to the intel setip . That took some time to create.

 

[Khato]

Took awhile to actually find some useful data for comparison - as usual AMD found a comparatively obscure benchmark to run that's not typically used by reviews and hence it's difficult to find numbers for. Luckily Google eventually came through - http://voz.vn/wp-content/uploads/2011/01/SnB-FF.png being the results themselves, from this page - http://null.vn/vbb4/showthread.php?...ield-v%E1%BB%9Bi-chi-ph%C3%AD-Lynnfield/page2 Note that it's in Vietnamese, so use of a translator is recommended.

So, i7 2600k HD 3000 scores a 845, i5 2500k HD 3000 scores a 702, while a radeon 5450 apparently on the same platform scores 746. So what? Being competitive with pathetic is still pathetic right?

It's unfortunate that AMD intentionally obscured the benchmark scores past a certain point - last clear screenshot is a score of 121. How does that help us? Well, it was pretty easy to find a youtube of someone running a radeon 5570 (on an AMD system even - http://www.youtube.com/watch?v=PjOZGtSuhK0) and going to the same point in the benchmark shows a score of approximately 184. So the Llano is at ~65% the performance of a 5570 running on a phenom II X3 at 2.8GHz. Now if we make an assumption that such would remain constant throughout the benchmark, then the score of 1914 that the 5570 gets would extrapolate to a score of 1259 for Llano... Oh, and by the same metric, the i7-2630QM system is on track for a score of 468, or 63% of what that review linked above obtained for an i5 2500k.

The point of all this being? AMD's demonstration is claiming that their solution is roughly 3 times as fast (their score is almost 3 times higher at the last point that both scores display simultaneously, 45 vs 121) whereas the extrapolated score is only 1.8x that of an i5 2500k, which has much faster CPU clocks than the i7-2630QM but the same IGP turbo clock of 1.1GHz. Feel free to draw your own conclusions as to why the i7-2630QM system is performing so poorly in the demonstration.

 

[Nemesis 1]

Originally Posted by Nemesis 1
Ya I can't wait to see AMDs version of thunderbolt that Apple is going to be using . I bet AMD spent billions in its development.

keep shoveling that dirt son, so how much money do you expect people to actually spend on disk to actually generate that amount of IO?

ps. i design high end enterprise networks for a living, including fully converged (data,san,voip) networks/datacenters. I have a very good idea how much disk that can generate that kind of IO costs both from low user count and high user count. one is really expensive and one is outrageous.

for SB based computer to reach llano GPU performance means discrete video card, that pushes up both power draw and cost.

you can really tell im currently in the documentation stage of this project, ie easily distracted by forums

Hay this was a reply to someone saying that Apple will run to AMD or should . Apple announced thunderbolt Apple won't be leaving intel any time soon . Your also not aware that Apple software just didn't run on X86 Intel had to make it work at considerable cost . Intel won't just give that work to apple so they can use AMd .

 

[itsmydamnation]

intel recompiled apples software for them...... a link for that? software is what apple do, encase you haven't noticed. I find the idea that they can't recompile there own software quite odd.

if apple see a target market and AMD can provide a solution that hits a perf/power/price/size point that intel can't you don't think apple will go with the AMD product?

 

[Nemesis 1]

Took awhile to actually find some useful data for comparison - as usual AMD found a comparatively obscure benchmark to run that's not typically used by reviews and hence it's difficult to find numbers for. Luckily Google eventually came through - http://voz.vn/wp-content/uploads/2011/01/SnB-FF.png being the results themselves, from this page - http://null.vn/vbb4/showthread.php?...c-mạnh-Bloomfield-với-chi-phí-Lynnfield/page2 Note that it's in Vietnamese, so use of a translator is recommended.

So, i7 2600k HD 3000 scores a 845, i5 2500k HD 3000 scores a 702, while a radeon 5450 apparently on the same platform scores 746. So what? Being competitive with pathetic is still pathetic right?

It's unfortunate that AMD intentionally obscured the benchmark scores past a certain point - last clear screenshot is a score of 121. How does that help us? Well, it was pretty easy to find a youtube of someone running a radeon 5570 (on an AMD system even - http://www.youtube.com/watch?v=PjOZGtSuhK0) and going to the same point in the benchmark shows a score of approximately 184. So the Llano is at ~65% the performance of a 5570 running on a phenom II X3 at 2.8GHz. Now if we make an assumption that such would remain constant throughout the benchmark, then the score of 1914 that the 5570 gets would extrapolate to a score of 1259 for Llano... Oh, and by the same metric, the i7-2630QM system is on track for a score of 468, or 63% of what that review linked above obtained for an i5 2500k.

The point of all this being? AMD's demonstration is claiming that their solution is roughly 3 times as fast (their score is almost 3 times higher at the last point that both scores display simultaneously, 45 vs 121) whereas the extrapolated score is only 1.8x that of an i5 2500k, which has much faster CPU clocks than the i7-2630QM but the same IGP turbo clock of 1.1GHz. Feel free to draw your own conclusions as to why the i7-2630QM system is performing so poorly in the demonstration.

Actually if ya go back and watch the video AMD had more than a few programs all running at the same time . The effects of which I not sure . But this videos is pure hype.

But those programms all running at same time may have played into what your reply informed us of . We just have to wait and see real results . I sure would like to know what gifts were exchanged between NV and Intel in that settlement . Intel payed 1.5 billion for something. It wasn't a punishment for intel booting NV chipsets.

 

[Mopetar]

But this videos is pure hype.

No, it's just a favorable scenario for AMD's chip. The chip is actually doing everything in that video. If anyone wants to read into it beyond that it's their own business, but it's a real video showcasing actual performance of a Llano CPU, which is actually quite good.

Unrealistic workload, most definitely. Pure hype, not be a long shot. I'm not sure why you seem to have so much trouble wrapping your mind around this concept. Llano appears to be quite good as a midrange gaming platform or for anything that can take advantage of its GPU.

 

[Nemesis 1]

intel recompiled apples software for them...... a link for that? software is what apple do, encase you haven't noticed. I find the idea that they can't recompile there own software quite odd.

if apple see a target market and AMD can provide a solution that hits a perf/power/price/size point that intel can't you don't think apple will go with the AMD product?

No link required I already posted the link along time ago . Test your search skills the info is here on this forum someware . And who said anything about a recompile I didn't . Your answeing questions and repling all over the place you should already know it wasn't a recompile . Maybe IDC will help you out . That dirt I was throwing . Burried you as far as I am concerned.

 

[IntelUser2000]

The point of all this being? AMD's demonstration is claiming that their solution is roughly 3 times as fast (their score is almost 3 times higher at the last point that both scores display simultaneously, 45 vs 121) whereas the extrapolated score is only 1.8x that of an i5 2500k, which has much faster CPU clocks than the i7-2630QM but the same IGP turbo clock of 1.1GHz. Feel free to draw your own conclusions as to why the i7-2630QM system is performing so poorly in the demonstration.

Your post was very interesting. I underclocked my system to the following:

2.4GHz, Turbo Mode off
Graphics Multiplier: 20, 1000MHz

The point that you were talking about I got 55.

 

[Nemesis 1]

No, it's just a favorable scenario for AMD's chip. The chip is actually doing everything in that video. If anyone wants to read into it beyond that it's their own business, but it's a real video showcasing actual performance of a Llano CPU, which is actually quite good.

Unrealistic workload, most definitely. Pure hype, not be a long shot. I'm not sure why you seem to have so much trouble wrapping your mind around this concept. Llano appears to be quite good as a midrange gaming platform or for anything that can take advantage of its GPU.

Really what game was that? I bet intel driver team was all over that game to get it running descent on its IGP . So AMD could do a comparsion with it . Actually I am surprized it even ran on HD graphics its such a popular game.

 

[Nemesis 1]

Your post was very interesting. I underclocked my system to the following:

2.4GHz, Turbo Mode off
Graphics Multiplier: 20, 1000MHz

The point that you were talking about I got 55.

re run the test with IGP locked at default 850mgz than tell us the score

 

[Khato]

Your post was very interesting. I underclocked my system to the following:

2.4GHz, Turbo Mode off
Graphics Multiplier: 20, 1000MHz

The point that you were talking about I got 55.

More importantly, was it having fun consistently pausing for a second or more at a time like in the demonstration? There's no question that Sandybridge is slower, but the stuttering in that video was very... odd and uncharacteristic of simple low frame rate.

Sadly, having a P67 based board means that I can't play with it myself =(

 

[itsmydamnation]

No link required I already posted the link along time ago . Test your search skills the info is here on this forum someware . And who said anything about a recompile I didn't . Your answeing questions and repling all over the place you should already know it wasn't a recompile . Maybe IDC will help you out .

when you go from one ISA to another you need to recompile, you might need to do work to make it recompile to another ISA but it still is a recompile. On the intel helped apple move i can't find anything in search of this forum or on the inet.

That dirt I was throwing . Burried you as far as I am concerned.

you should become a comedian, i will let other people make up there own minds but to me all you have done is wildly made accusations, been completely dismissive without being able to logically say why, never actual address any counter points to your wild rants.

Given that the 18watt apu can match a mobile SB in GPU performance what do you think a 60watt APU is gonna do, then your dismissive when it does.

basically i know im being trolled.........yay :thumbsup:

 

[Nemesis 1]

Ya i just downloaded the FF benchmark went to load the game and I have a missing dill. So I will maybe load it on to the gamer tomorrow see how that turns out. I wanted to see that slide show on this old p4c with x1950 pro and thats what it should look like a slide show according to the AMD video slower than a slide show on this Pc.

 

[Nemesis 1]

when you go from one ISA to another you need to recompile, you might need to do work to make it recompile to another ISA but it still is a recompile. On the intel helped apple move i can't find anything in search of this forum or on the inet.



you should become a comedian, i will let other people make up there own minds but to me all you have done is wildly made accusations, been completely dismissive without being able to logically say why, never actual address any counter points to your wild rants.

Given that the 18watt apu can match a mobile SB in GPU performance what do you think a 60watt APU is gonna do, then your dismissive when it does.

basically i know im being trolled.........yay :thumbsup:

Try google search for when apple left IBM for Intel x86. What accusations are wild , We have seen this type of thing from AMD befor so its not wild but the norm.