• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Trinity review

Page 11 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
mrcmtl said:
Llano's Turbo Core is far from useless. It is not the Turbo that existed in Thuban but the improved version similar to the one in BD. In purely single threaded apps, Llano is in turbo state more than 50% of the time.
Click to expand...

Yes but it barely works in anything but 1 core. You can see from the performance. Once you figure out stable clocks of Turbo, there's no difference between desktop and laptop chips in performance/clk. If it looks like there is, its because the clocks on laptops vary so much.

Pelov: Here's a clock for clock comparison between Llano, Trinity, and Bulldozer: http://www.4gamer.net/games/133/G013372/20120514010/

Although its in synthetic benchmarks, the results are still interesting.
 
IntelUser2000 said:
Here's a clock for clock comparison between Llano, Trinity, and Bulldozer: http://www.4gamer.net/games/133/G013372/20120514010/

Although its in synthetic benchmarks, the results are still interesting.
Click to expand...

Nice find, kudos to your review prowling skills. Granted there's not much of a conclusion to be drawn from the results as it's unknown how the difference in L3 cache affects results. But it is telling that the A10-4600M is ahead of the FX-4100 by a non-trivial margin in one test (CPU Zlib) and the translated text makes it sound like that may be due more to the Trinity system having an SSD. Here's to hoping that a certain domestic review site rises to the challenge and does a similar investigation! I can hope, right?
 
Some pretty solid gains from OpenCL in a non-public build of Handbrake/x264 that supports GPU acceleration. Nice to see more useful software leveraging the GPU.

46687.png


http://www.anandtech.com/show/5835/testing-opencl-accelerated-handbrakex264-with-amds-trinity-apu

And sort of off topic, but has anyone messed around with WinZip 16.5's OpenCL acceleration on an AMD Brazos system? Installed the trial version on a C-60 netbook and didn't really see any difference with OpenCL enabled, wasn't much/any faster and GPU utilization doesn't get that high, looks like it's barely using the GPU compared to w/out OpenCL enabled. Wonder if the CPU or something else is bottlenecking me. Or maybe it's a driver issue, might try updating them.
 
frostedflakes said:
Wonder if the CPU or something else is bottlenecking me. Or maybe it's a driver issue, might try updating them.
Click to expand...

I suppose 80SPs are not enough to outperform the CPU. They need 400SPs with many more ROPs and TMUs to outperform the CPU by 2x. The iGPU in Brazos has 1/5th the compute power and corresponding decrease in everything else, and the memory BW is lower too. Even if the CPU is slower, its not 1/5th Llano: http://www.anandtech.com/bench/Product/385?vs=345
 
Kalessian said:
Nobody games on a laptop... should be nobody SHOULD game on a laptop, but tons of my friends still use core2 based laptops to play mmos/rpgs/rts/mobas. It makes me cringe. I bet any of them would love a $600 trinity laptop upgrade.
Click to expand...

My grandson still games on an Asus core 2 duo laptop from 3 years ago that runs a gtx260m. I bet it is still faster in games than ivy or trinity, although it will run hotter. Point is igps still have a long way to go in gaming. Point 2 is I think there is a place for gaming laptops. You still need a discrete gpu though IMO.
 
frozentundra123456 said:
My grandson still games on an Asus core 2 duo laptop from 3 years ago that runs a gtx260m. I bet it is still faster in games than ivy or trinity, although it will run hotter. Point is igps still have a long way to go in gaming. Point 2 is I think there is a place for gaming laptops. You still need a discrete gpu though IMO.
Click to expand...

lol.....your gtx260 uses 75Watts alone, more than double of trinity max...

amd should realesa a 60 watt trinity 😎
 
The i7-3720QM is simply not in the same price bracket at the A10. Go price check any notebook with a i7-3720QM. They are all seriously like $1200-$1800. I'd be surprised to find one on slickdeals for under $900 any time this summer.

On the other hand I would be shocked to not be able to find an A10 notebook for $600 in the next few months.

I think it is safe to assume that if a i7-3720QM is 20% slower in games then a $600 i3 ib is going to be 40-50% slower.
 
sm625 said:
The i7-3720QM is simply not in the same price bracket at the A10. Go price check any notebook with a i7-3720QM. They are all seriously like $1200-$1800. I'd be surprised to find one on slickdeals for under $900 any time this summer.

On the other hand I would be shocked to not be able to find an A10 notebook for $600 in the next few months.

I think it is safe to assume that if a i7-3720QM is 20% slower in games then a $600 i3 ib is going to be 40-50% slower.
Click to expand...

Interesting assumption given the mobile i7, i5, and i3 use/will use HD 4000 graphics. There's also no CPU bottleneck. There will probably be some difference because of lower IGP frequency, which would perhaps put it at 5% slower than the HD 4000 in the i7. Also, the A8 has a gimped IGP portion and will probably be very comparable to equal to the HD 4000.

How did you arrive at your conclusion?
 
IntelUser2000 said:
Here's a clock for clock comparison between Llano, Trinity, and Bulldozer: http://www.4gamer.net/games/133/G013372/20120514010/

Although its in synthetic benchmarks, the results are still interesting.
Click to expand...

Now that's surprising! I read it this morning as I was leaving the house and haven't had time to comment.

The CPU Queens benchmark I know uses very very little cache -- in fact I believe it fits into L1 -- so the L3 shouldn't make any difference there. It does test branch prediction and it seems to be flat... I'm not sure about the other benchmarks and how/if they utilize the L3 cache, but that's still inexcusable to see they've essentially flatlined in their IPC improvements. If Vishera doesn't clock to 5ghz on turbo AMD is screwed.

I don't think the benchmarks take anything away from Trinity, though. Considering the other improvements alongside the CPU improvements its actually still a good chip. It just doesn't bode well for Vishera :/
 
Intel17 said:
Yeah. Trinity APU for laptops will be one hell of a value...
Click to expand...

Not really, no.

For $500 you'll probably find an Ivy Core i3 laptop. Those will have HD 4000 graphics, which will be about equal to the HD 7640G featured in the A8 APU. The i3 will smash the A8 in CPU performance and be on equal terms in IGP, and battery life will be comparable.

For $600 you'll probably be able to find either a Core i3 laptop + dGPU or Core i5. That's the same for an A10, and an i3+ switchable dGPU (7670M) would be much faster in everything. A Core i5 would leave the A10 in the dust in CPU performance.

Please explain where this so-called value is because I'm failing to see it compared to Ivy Bridge.
 
LOL_Wut_Axel said:
Interesting assumption given the mobile i7, i5, and i3 use/will use HD 4000 graphics [This is wrong as we know nothing about the i3, not even a release date]. There's also no CPU bottleneck [There is no CPU bottleneck in Trinity either. Take a look at the Starcraft numbers]. There will probably be some difference because of lower frequency, which would perhaps put it at 5% slower than the HD 4000 in the i7 [Where are you getting this? Intel shares it's cache between CPU and GPU and considering the i5's haven't even been tested and will have half the L3 of the i7 that's been tested it's likely to be more than 5%].

How did you arrive at your conclusion?
Click to expand...

I suppose we can ask you the same thing D:
 
This is wrong as we know nothing about the i3, not even a release date
Click to expand...

It is not. The Ivy mobile i3 will have HD 4000 graphics, just like the Sandy mobile i3 had HD 3000 graphics. The difference will be their Turbo frequencies, which should mean a small ~5% difference (in some cases perhaps 10%).

There is no CPU bottleneck in Trinity either. Take a look at the Starcraft numbers
Click to expand...

I never said there was. I was talking about Ivy Bridge, not Trinity. However, if you do add a dGPU to Trinity, a CPU bottleneck will definitely appear.

Where are you getting this? Intel shares it's cache between CPU and GPU and considering the i5's haven't even been tested and will have half the L3 of the i7 that's been tested it's likely to be more than 5%
Click to expand...

That makes an insignificant difference. What matters to IGPs is memory bandwidth, and Intel has heaps more than AMD at their disposal. 3-4MB of L3 shared with the CPU is more than enough for the IGP.

Nice try, though.
 
The L3 can make a considerable difference, particularly in gaming. Take a look:

crysis_1024.gif

Mind you that's a CPU only and it's losing to a Phenom with L3 and 200mhz deficit in clock speeds.

It's partly the reason why that Trinity CPU would look mediocre in gaming compared to an SB/IB alternative with the same discrete GPUs. That 30% gap in single-threaded performance also doesn't help 😛

I'm not sure where you're getting your IB i3 info (probably the same place you get all of your other info which is quite a mystery to a lot of us), but from what I've seen there hasn't been a release date, never mind clock speeds or what graphics it will have. The only thing we've gotten for the i3 are more delays, which kind of sucks given that I'm hoping to choose between an i3 and a Trinity APU 🙁
 
Last edited:
pelov said:
The L3 can make a considerable difference, particularly in gaming. Take a look:

crysis_1024.gif


It's partly the reason why that Trinity CPU would look mediocre in gaming compared to an SB/IB alternative with the same discrete GPUs. That 30% gap in single-threaded performance also doesn't help 😛

I'm not sure where you're getting your IB i3 info (probably the same place you get all of your other info which is quite a mystery to a lot of us), but from what I've seen there hasn't been a release date, never mind clock speeds or what graphics it will have. The only thing we've gotten for the i3 are more delays, which kind of sucks given that I'm hoping to choose between an i3 and a Trinity APU 🙁
Click to expand...

You do realize there's a difference between the effectiveness and efficiency of AMD's and Intel's L3 cache architecture, right? Intel's architecture makes much more effective use of cache by making it faster, which matters more than putting gobs of it (which is what AMD does with FX). Again: having faster L3 cache is more important than having a huge amount of it. That's the exact reason why there's no difference in gaming at the same clocks between a 2500K/3570K and a 2600K/3770K. If you look at the graph posted above, you'll notice the i3 and i5 (Clarkdale) are the fastest even though they have half the L3 cache of the Phenom II. That again reinforces my point about fast L3 cache>>>lots of slow L3 cache.

Having high memory bandwidth is also very important, and that's an area where Intel have AMD squarely beat. The HD 4000 graphics have enough memory bandwidth to go around.

And the mobile i3 will have HD 4000 graphics, again. It's the standard desktop version (3xx0) which will use HD 2500 graphics.
 
LOL_Wut_Axel said:
And the mobile i3 will have HD 4000 graphics, again. It's the standard desktop version (3xx0) which will use HD 2500 graphics.
Click to expand...

Where are you getting this? I'd like to know. Intel hasn't released anything at all about the i3's other than a very very rough guesstimated release date of "Summer."
 
pelov said:
Where are you getting this? I'd like to know. Intel hasn't released anything at all about the i3's other than a very very rough guesstimated release date of "Summer."
Click to expand...

Because I know a lot about this industry.

The mobile Ivy Bridge Core i3 will have HD 4000 graphics and the desktop version will have HD 2500 graphics except for the 3xxx5 versions. Intel is following exactly the same hierarchy with Ivy Bridge that they did with Sandy Bridge when it comes to product placement. As a result of that, you'll see HD 4000 graphics in the mobile i3, i5, and i7 with the differences being Turbo frequencies. This is the same as Sandy Bridge.

Again, same product placement:
HD 4000>HD 2500>HD Graphics
HD 3000>HD 2000>HD Graphics

This seems rather simple. You should understand it by now.

Also, you should stop trying to deflect. You missed the most important part (the whole part about L3 cache) to focus on something else entirely.
 
So, just to clarify, you're saying this because you know a lot about this industry?

Sort of like you said that people care more about CPU performance in mobile and yet people were and are still buying Llanos at higher rates than Intel alternatives, right?

I'm asking because your statements have a pretty solid history of being wrong. Do you have any legitimate proof that the rest of us can use?
 
Please be sure to post in the Hot Deals forum when you see an i3+7670m for $600 or less, especially if it has 14 or 13.3 inch screen.

*New with minimum 1 year warranty, don't think refurbished or outlet would be a deal at $500-600.

LOL_Wut_Axel said:
For $600 you'll probably be able to find either a Core i3 laptop + dGPU or Core i5. That's the same for an A10, and an i3+ switchable dGPU (7670M) would be much faster in everything. A Core i5 would leave the A10 in the dust in CPU performance.
Click to expand...
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top