Mobile Intel Ivy Bridge vs AMD Trinity?

[AtenRa]

And I've heard multiple AMD fans say CPU is not important to graphics, only turn around and say it is, when argument merits it. There's no difference in graphics between the 35W and 45W Llano chips, only the CPU. For Sandy Bridge/Ivy Bridge, using a lower end CPU is sometimes favorable because of the TDP sharing.

CPU is not important unless you are CPU bottleneck. At 1366x768 and Medium IQ settings the Llano CPU most probable bottlenecks the iGPU.

If the CPU is the bottleneck, then a faster CPU Llano will perform better.

 

[Abwx]

CPU is not important unless you are CPU bottleneck. At 1366x768 and Medium IQ settings the Llano CPU most probable bottlenecks the iGPU.

If the CPU is the bottleneck, then a faster CPU Llano will perform better.

He indeed agree with this point when it suits his demonstration...

Also, improvements between CPU and GPU can't be compared equally. 10% on CPU might be just as big as 30-40% on GPU. The former speeds up everything while latter only helps on 3D graphics-like

 

[Arzachel]

Llano is like Clarkdale. Maybe the variation is even less then Clarkdale. The Turbo Mode on Llano chips barely work, and doesn't have GPU Turbo at all.

I don't even mean turbo, bumping up the CPU clock 25% higher would make the Llano system perform quite a bit better. That is why I find it weird that they didn't use a A8 3550mx, but I guess reviewers have to make due with what they get sent.

 

[IntelUser2000]

He indeed agree with this point when it suits his demonstration...

Way to relate something that's not related at all. The 10% gain in CPU benchmarks can't be compared with 30-40% on GPU benchmarks. Because the latter already gets minimal fps increase using CPU since GPU is bottlenecking it.

Do you seriously not get this? It's called a bottleneck, man. A CPU that might get 10% in CPU-bottlenecked applications might only get 3-4% faster in games.

 

[Abwx]

Way to relate something that's not related at all. The 10% gain in CPU benchmarks can't be compared with 30-40% on GPU benchmarks. Because the latter already gets minimal fps increase using CPU since GPU is bottlenecking it.

Do you seriously not get this? It's called a bottleneck, man. A CPU that might get 10% in CPU-bottlenecked applications might only get 3-4% faster in games.

Agree on this point as well as for the majority of yours posts ,
the only dispute i made was about comparing vastly different
CPU and then draw conclusions out of discutable benchmark
set ups as this shamefull Anand s one in a review whose conditions
have been very obviously been set up by Intel itself , albeit
all the sites that were delivered SKUs has to follow those
twisted guidelines.

As for the eventual bottlenecks, you will surely agree that a 1.6g CPU will hardly be bottlenecked by a 400 SP radeon , it s rather the contrary.

 

[IntelUser2000]

Memory impact of Llano. http://forum.notebookreview.com/gam...g-benchmarked-various-ram-configurations.html

Difference between Asynchronous DDR3-1333 to Synchronous DDR3-1600 is 9%. Gap between DDR3-1600 and DDR3-1333 when both are synchronous are even less at 6%. Memory impact is huge for desktop Llanos, not as much for laptop.

CPU difference on gaming: http://www.tomshardware.com/reviews/a8-3500m-llano-apu,2959-8.html

Phenom II X4 running at 2.4GHz is 0-1% faster than one running at 1.5GHz. Trinity is supposed to be 56% faster in 3DMark11, the gains are all due to GPU.

 

[itsmydamnation]

Way to relate something that's not related at all. The 10% gain in CPU benchmarks can't be compared with 30-40% on GPU benchmarks. Because the latter already gets minimal fps increase using CPU since GPU is bottlenecking it.

Do you seriously not get this? It's called a bottleneck, man. A CPU that might get 10% in CPU-bottlenecked applications might only get 3-4% faster in games.

nice to see your pretty much flat out lying there :thumbsdown:. everything so far, AMD, Leaks, the posts/benchmarks on Semi accurate all point to much higher then 10% improvement, hell some CPU benchmarks are seeing 50% ( an outlier, but still happened). Hell with the same TDP the benifit from the resonance clock mesh should almost hit your 10% performance increase.

That completely ignores improvements to the process over time and changes to cores and the much improved turbo. doesn't matter if you compare it to STARS or bulldozer pile-driver SOC's in CPU out perform by more then 10%.


for reference i wonder what your posts during the pre release for bulldozer were like?

 

[IntelUser2000]

nice to see your pretty much flat out lying there :thumbsdown:.

BTW, that was about Ivy Bridge, idiot.

For Bulldozer I estimated that it would be 10-20% faster in multi-threaded applications than 2600K based on AMD's earlier claims.

 

[pelov]

Worth a read

Looks like they've done quite a bit of fiddling with the Bulldozer cores. I'm wondering how low they managed to get the L2 latency as that was easily one of the bigger issues with Bulldozer. The reduced leakage is a welcome sign as well.

The AMD claims on Bulldozer were the biggest load of horseshit I've ever seen. JF immediately disappeared after its release in order to not live up to the shame. Frankly that was by far AMD's biggest mistake with Bulldozer. By overstating their improvements they built up a lot of hype and then never managed to live up to expectations.

The difference this time around seems to be that the leaked benchmarks and figures point to AMD's claims being on target. Prior to Bulldozer there were leaked figures and signs, numerous delays that pointed to some serious issues. This time around there's been a far more consistent picture painted and the picture looks pretty good for AMD. The real question is the price. They fucked it up bad with Bulldozer and it's only now that BD is dropping in price to where you can make the argument that it makes sense but it's taken an IB release and numerous SB price drops as well and now BD arguably looks even less attractive than it did at launch. Trinity's die size is slightly bigger than Llanos but apparently no yield issues so let's hope Read was serious about beating Intel on price.

 

[Don Karnage]

Worth a read

Looks like they've done quite a bit of fiddling with the Bulldozer cores. I'm wondering how low they managed to get the L2 latency as that was easily one of the bigger issues with Bulldozer. The reduced leakage is a welcome sign as well.

The AMD claims on Bulldozer were the biggest load of horseshit I've ever seen. JF immediately disappeared after its release in order to not live up to the shame. Frankly that was by far AMD's biggest mistake with Bulldozer. By overstating their improvements they built up a lot of hype and then never managed to live up to expectations.

The difference this time around seems to be that the leaked benchmarks and figures point to AMD's claims being on target. Prior to Bulldozer there were leaked figures and signs, numerous delays that pointed to some serious issues. This time around there's been a far more consistent picture painted and the picture looks pretty good for AMD. The real question is the price. They fucked it up bad with Bulldozer and it's only now that BD is dropping in price to where you can make the argument that it makes sense but it's taken an IB release and numerous SB price drops as well and now BD arguably looks even less attractive than it did at launch. Trinity's die size is slightly bigger than Llanos but apparently no yield issues so let's hope Read was serious about beating Intel on price.

Agreed. The 8120s price is very attractive and will only get better

 

[Olikan]

did i said? piledriver is a page long of small tweaks over bulldozer....pro tip, there is more changes in vishera (more than just L3 cache)