AMD A10-4600M Review: Mobile Trinity Gets Tested

[zooted]

I am glad to see the increase in ipc over previous generation. This is a good step. Interesting to see a fully functional desktop version of Trinity and especially Piledriver.

 

[kartu]

[citation][nom]SignMeUpSammy[/nom]Based on this, gaming is much better than old i5, but everything else including application performance is still better on the old Sandy architecture. I'm not really sure why I would buy a Trinity other than for a casual gaming laptop. Unfortunately, budget says that my laptops have to be used for business first, play time later.[/citation]
Exactly in which app typically used at home you'd be losing with Trinity? (let alone the fact, that AMD's motherboard + CPU cost much less) Your excel will be milliseconds faster? Uhm, browsers would render pages milliseconds faster?

Most of the peformance bottlnecked tasks on PC come down to games and video encoding. But in case of the latter, I don't care whether PC will finish at 2:00 at night or at 2:40.

 

[cleeve]

Those are some interesting notes, Palladin.

As far as setting thread affinity, I've never heard of that and AMD certainly hasn't mentioned it, but I'm sure going to try it! If it actually works, I'll probably write up a follow-up review. And if that's the case it's insane that AMD doesn't know/tell reviewers/address this.

Where memory is concerned, Trinity reported 1600 MHz, yes the bandwidth showed up low in Sandra but I've seen Sandra misreport bandwidth so many times now that I don't even trust it. But I will follow up with some other memory tests to make sure.

 

[blazorthon]

[citation][nom]Cleeve[/nom]Those are some interesting notes.As far as setting thread affinity, I've never heard of that and AMD certainly hasn't mentioned it, but I'm sure going to try it! If it actually works, I'll probably write up a follow-up review. And if that's the case it's insane that AMD doesn't know/tell reviewers/address this.As far as memory, Trinity reported 1600 MHz, yes the bandwidth showed up low in Sandra but I've seen Sandra misreport bandwidth so many times now that I don't even trust it. But I will follow up with some other memory tests to make sure.[/citation]

Is there a more consistently accurate way to measure the memory bandwidth?

EDIT: Or at the least, is there a source that confirms whether or not that the memory controller in Trinity is an improvement over Llano?

 

[badtaylorx]

@caedenv no dif between a 5yo computer and a new i7+ discrete graphics??? either you have that 5/yo desktop and are in complete denile, or a troll....i have both,,, there is a HUGE difference.

on another note,,, this does bode well for amd 990fx owners...i may have to dust off the bulldozer!!!

 

[blazorthon]

[citation][nom]badtaylorx[/nom]@caedenv no dif between a 5yo computer and a new i7+ discrete graphics??? either you have that 5/yo desktop and are in complete denile, or a troll....i have both,,, there is a HUGE difference.on another note,,, this does bode well for amd 990fx owners...i may have to dust off the bulldozer!!![/citation]

If Piledriver is as much of an improvement as it seems to be, then Piledriver CPUs might be pretty good. CPUs do't need to be super fast, but Bulldozer was too far of a step back. Piledriver looks like it will at least beat Phenom II and hopefully Core 2 (I'm really hoping for it at least coming close to Nehalem, but that might be wishful thinking), making its CPUs much more viable than Bulldozer CPUs for most workloads.

 

[blazorthon]

[citation][nom]tourist[/nom]Just wondering how this will all play out when ddr4 arrives next year, it would seem amd is setting themselves up to take full advantage of the increased bandwidth.[/citation]

Maybe... However, I'm not sure. If AMD support it in Trinity's successor, then it's a recipe for diminishing the memory bandwidth bottleneck greatly. If the next APUs are 28nm with GCN cores, then it could be pretty important to have more memory bandwidth, even more so than it is today, if Trinity's successor beats Trinity better than Trinity beats Llano in graphics performance. However, I'm pretty sure that DDR4 won't see much, if any, usage in the consumer markets until 2014. If I remember correctly, only servers will see DDR4 in 2013, so it might not be until two APU generations after Trinity that it sees usage in the APUs.

 

[bobafert]

[citation][nom]Wisecracker[/nom]Some enterprising youts can start extrapolating and charting.[/citation]

not interested in what AMD writes in their press breifs. I want to see the numbers. I read the article and saw all the charts and I want more.

 

[Blink]

I'm looking forward to seeing Trinity vs HD4000 when Intel releases a comparably priced model. I've been chomping at the bits to build a HTPC that will also play not so ancient titles. Unfortunately my finances may not be ready to make luxury purposes until Steamroller.

 

[sonofliberty08]

im waiting for brazos 2.0

 

[The_Trutherizer]

[citation][nom]blazorthon[/nom]HD 4000 on the top end models against mid-range and low end APUs... Yeah, there's a reason that Tom's didn't compare them much. They are for different markets. In fact, this has been said several times in the comments already. It's common sense, not madness, to compare something to another in the same price range.[/citation]

Yet Andandtech was at least curious and they found that Trinity kicked gaming butt vs the HD4000 CPUs. Of course processing power was less than equal, but it is worth noting.

 

[Guest]

APU trinity offers enormous potential. The test bench for the dimostrano.Molti software are obsolete and underestimate the performance reali.Oggi I compared the hardware of my notebook with the new APU and other competitive products using Pass Mark.Dall 'analysis of these tests can tell you that APU trinity that are truly exceptional ... especially when compared with the previous version Llano

 

[wiyosaya]

[citation][nom]notyetapu[/nom]neoverdugojust a side note, what you described is not an apu, it's a cpu with on die gfx. AMD's apu have not hit their full stride yet, once we have mature implementation of gpu assisted processing (opencl directcl et al) then the disparity may become significantly less, AMD strategy was always to leverage the massive computing power of the gfx core to bolster cpu performance in areas other than gaming unfortunately there was a fragmentation of the market with competing standards, once all that mess gets sorted out AMD can really flex the power of the apu[/citation]
True, and AMD does have the GPU advantage over Intel.

However, just because you run a task on the GPU does not mean it will run faster than it does on the CPU. The more the task lends itself to parallel processing, the faster it will run on the GPU in comparison to the CPU. If the task does not lend itself to parallel processing, then there may be no gain in speed between running it on the GPU and the CPU.

Parallel processing, which is what GPUs excel in doing, is not necessarily an obvious implementation.

 

[Avro Arrow]

[citation][nom]jacobdrj[/nom]I have found that when you have 3 or more threads, the pinch points are not the CPU power or the CPU thread counts. The pinch points are A) the amount of memory you can address and B) the amount/speed of your virtual memory.I just put 16 GB of RAM on an A8-3500 laptop... Turned off the virtual memory... Runs like a dream, even on Firefox, and even with 20 open tabs... Assuming you use a 64 bit build...[/citation]
And just how the hell is anyone supposed to use 16GB of RAM with a 32 bit build? LMAO!!!!
[citation][nom]blazorthon[/nom]Yeah, that could launch Trinity far above all other IGPs (although I think that 1GB would be better for the top Trinity APUs, especially on the desktop side) in performance.[/citation]
I seriously doubt that. No video card that is less powerful than a desktop ATi Radeon HD 4870 is powerful enough to use more than 512MB which is why most of the nVdia GTX 2xx series only had 768MB on board. I remember reading an article where the author only managed to get the HD 4870 to use more than 512MB in a very specific instance of Crysis.

 

[blazorthon]

[citation][nom]Avro Arrow[/nom]And just how the hell is anyone supposed to use 16GB of RAM with a 32 bit build? LMAO!!!!I seriously doubt that. No video card that is less powerful than a desktop ATi Radeon HD 4870 is powerful enough to use more than 512MB which is why most of the nVdia GTX 2xx series only had 768MB on board. I remember reading an article where the author only managed to get the HD 4870 to use more than 512MB in a very specific instance of Crysis.[/citation]

... A big part of the IGP here is for it to manage in GPGPU workloads to make up for the inferior CPU performance compared to it's competition. Having 1GB instead of 512MB would help that. Also, the 4870 is not a DX11 GPU, among other deficiencies. it does not need to use as much memory. A better card to test memory utilization would be using the Radeon 6770 or at least the 5770, they have similar performance to the 4870, but are more updated GPUs. Also, I bet if you tested using a game newer than Crysis, such as Metro 2033 or BF3, it will use more memory than Crysis did.

Also, jacodrj specified that his/her post only applies if you use a 64 bit OS in his/her last sentence.

 

[lott11]

Tom's you miss the boat again.
Directs 9 again!!
So what happens to Directs 11, and where is the ivy.?
Do a real comparative test.
If there is real test just do it, do not sugar code it one way or the other.
We are all grownups, be real journalist.
And really cover all test dual monitor test, multi task loads, and really load the test.
A real world test lets say multi desktop load like Windows 8 and Ubuntu 12.4 .
where you have 4 desks top running at the same time doing multiple applications all in real time.
Yes windows 8 is beta, but that is what people are going to use in the future, and most people do multi task.
That is a real test, that is what real people do in the real world, not stats.
A bench mark that those little, but to promote but a bench march software.
All that we hear is that later on this year when it comes out, we do those test.
You are suppose be journalist, do real investigation show us it's weakness and strengths.
That is all we ask for, nothing less & no more.

 

[blazorthon]

[citation][nom]lott11[/nom]Tom's you miss the boat again.Directs 9 again!!So what happens to Directs 11, and where is the ivy.?Do a real comparative test. If there is real test just do it, do not sugar code it one way or the other.We are all grownups, be real journalist. And really cover all test dual monitor test, multi task loads, and really load the test.A real world test lets say multi desktop load like Windows 8 and Ubuntu 12.4 .where you have 4 desks top running at the same time doing multiple applications all in real time.Yes windows 8 is beta, but that is what people are going to use in the future, and most people do multi task. That is a real test, that is what real people do in the real world, not stats. A bench mark that those little, but to promote but a bench march software.All that we hear is that later on this year when it comes out, we do those test. You are suppose be journalist, do real investigation show us it's weakness and strengths.That is all we ask for, nothing less & no more.[/citation]

HD 3000 and 2000 don't support DX11, so it can't be used for the benchmarks. Windows 8 previews are not guaranteed to act exactly how the full release versions will, so including the previews would be irrelevant and would not change anything. Including Ubuntu would be pointless for most people and they were probably time-constrained, so they did not have time to test Ubuntu. Furthermore, Ubuntu might not even have driver support for any of the next gen IGPs yet.

Moving on, the HD 4000 in different Intel CPUs performs differently, so including the high end Ivy Bridge CPUs would have not only been comparing two parts from different markets and budget ranges, but would also probably not be representative of the performance that we would get from lower end CPUs with HD 4000.

 

[palladin9479]

Those are some interesting notes, Palladin.

As far as setting thread affinity, I've never heard of that and AMD certainly hasn't mentioned it, but I'm sure going to try it! If it actually works, I'll probably write up a follow-up review. And if that's the case it's insane that AMD doesn't know/tell reviewers/address this.

Where memory is concerned, Trinity reported 1600 MHz, yes the bandwidth showed up low in Sandra but I've seen Sandra misreport bandwidth so many times now that I don't even trust it. But I will follow up with some other memory tests to make sure.

Not AMD / Intel's fault really, there is nothing they can do in hardware to remedy the issue.

The problem is how Windows does scheduling, no thread is given 100% of a CPU's time. Threads are routinely interrupted so that the OS can query the CPU and possibly task switch another thread onto it. When Windows goes to resume execution on the thread it doesn't always put it back where it got it, where it puts it is based on the CPU utilization of each core. If the CPU is powering down / down clocking under-used cores then Windows will see them as under-used cores and try to assign that thread to them. That cause's them to spike back up. Clocking up / down is determined by a CPU utilization threshold vs a time period. Llano's were factory set to only clock up if there was > 60% utilization for 500ms, and only clock down if there is <40% utilization for more then 2000ms. TB only activates if two or more cores are at their lower P states, I haven't had time to figure out exactly where the threshold was on Llano. I do know that if three cores are running at their P0 state then TB will never activate.

Realizing the logic behind boosting you can see why we rarely see it activate on anything in Windows, even if it's a single threaded application. Windows is constantly moving it around which prevents the CPU from ever properly down clocking / boosting. Forcing processor affinity is a way to prevent windows from moving it around and thus allows for the conditions where TB can activate. I've done it multiple times on my A8-3550MX.

 

[jacobdrj]

[citation][nom]Avro Arrow[/nom]And just how the hell is anyone supposed to use 16GB of RAM with a 32 bit build? LMAO!!!!I seriously doubt that. No video card that is less powerful than a desktop ATi Radeon HD 4870 is powerful enough to use more than 512MB which is why most of the nVdia GTX 2xx series only had 768MB on board. I remember reading an article where the author only managed to get the HD 4870 to use more than 512MB in a very specific instance of Crysis.[/citation]

You're not. Some new computers are sold with a 32 bit os, but have 4 gb ram: They can't use all of it... What was your your point?

 

[lott11]

blazorthon
HD 3000 and 2000 don't support DX11, so it can't be used for the benchmarks. Windows 8 previews are not guaranteed to act exactly how the full release versions will, so including the previews would be irrelevant and would not change anything. Including Ubuntu would be pointless for most people and they were probably time-constrained, so they did not have time to test Ubuntu. Furthermore, Ubuntu might not even have driver support for any of the next gen IGPs yet.

Moving on, the HD 4000 in different Intel CPUs performs differently, so including the high end Ivy Bridge CPUs would have not only been comparing two parts from different markets and budget ranges, but would also probably not be representative of the performance that we would get from lower end CPUs with HD 4000.

Lott
That is the point.
In one instance you are comparing cost, then when suets you compare performance.
Intel dose not support Directs 11 with out discreet graphics, but AMD dose not support 9 with out discreet graphics.
That is what you would cover a level comparison, you would do both test.
For most people they are trying to get the most for the least, then there enthusiast, coders & gamers.
I am both both type of people an enthusiast and budget builder.
For enthusiast for coding, and budget builder for the rest of my home PC'S.
I replace all PC's and laptop every 3 years, and donate my older unit with Linux to School or Friends.
So why is the comparison so out in the Ball park?
You compared both AMD APU, do the same with Intel's no matter the cost you have done it before.
That is what you are comparing is it not.
And if windows 8 can not be used, then do a multi task on windows 7.
It Simple do a multi task run, and just do cascade desk top.
Would it be so complicated.
What is done on what time, and what is done first.
Ho that would not take that long to do, a simple macro.
Multi tasking is not complicated, that is a real world test.

 

[Avro Arrow]

You know, I just realised something. Looking at the above chart, Trinity is about 50% better than Llano but it's also clocked 50% higher. Is it just me or is Trinity's IPC roughly the same as Llano?

 

[blazorthon]

[citation][nom]lott11[/nom]blazorthonHD 3000 and 2000 don't support DX11, so it can't be used for the benchmarks. Windows 8 previews are not guaranteed to act exactly how the full release versions will, so including the previews would be irrelevant and would not change anything. Including Ubuntu would be pointless for most people and they were probably time-constrained, so they did not have time to test Ubuntu. Furthermore, Ubuntu might not even have driver support for any of the next gen IGPs yet.Moving on, the HD 4000 in different Intel CPUs performs differently, so including the high end Ivy Bridge CPUs would have not only been comparing two parts from different markets and budget ranges, but would also probably not be representative of the performance that we would get from lower end CPUs with HD 4000.LottThat is the point. In one instance you are comparing cost, then when suets you compare performance.Intel dose not support Directs 11 with out discreet graphics, but AMD dose not support 9 with out discreet graphics.That is what you would cover a level comparison, you would do both test.For most people they are trying to get the most for the least, then there enthusiast, coders & gamers.I am both both type of people an enthusiast and budget builder.For enthusiast for coding, and budget builder for the rest of my home PC'S.I replace all PC's and laptop every 3 years, and donate my older unit with Linux to School or Friends. So why is the comparison so out in the Ball park?You compared both AMD APU, do the same with Intel's no matter the cost you have done it before.That is what you are comparing is it not.And if windows 8 can not be used, then do a multi task on windows 7.It Simple do a multi task run, and just do cascade desk top. Would it be so complicated.What is done on what time, and what is done first.Ho that would not take that long to do, a simple macro. Multi tasking is not complicated, that is a real world test.[/citation]

AMD supports DX9 on all of their APUs in addition to DX10 and DX 11. Intel supports DX11 on their HD 4000 and up to DX10.1 on their HD 2000 and HD 3000 IGPs. You clearly don't know what you're talking about if you think otherwise because that is simply how it is. I'm not even sure what it is you are asking for at the end of your comment.

Trinity is compared against Llano because Trinity is Llano's successor. It would be ridiculous to not compare it to Llano. Tom's decided to not compare the much more expensive Intel CPUs with HD 4000 against the Trinity APUs because they are in two distinct budget ranges, IE they don't compete with each other. If you want to see them compared, then just check the Anand article.

 

[A Bad Day]

[citation][nom]palladin9479[/nom]Not AMD / Intel's fault really, there is nothing they can do in hardware to remedy the issue.The problem is how Windows does scheduling, no thread is given 100% of a CPU's time. Threads are routinely interrupted so that the OS can query the CPU and possibly task switch another thread onto it. When Windows goes to resume execution on the thread it doesn't always put it back where it got it, where it puts it is based on the CPU utilization of each core. If the CPU is powering down / down clocking under-used cores then Windows will see them as under-used cores and try to assign that thread to them. That cause's them to spike back up. Clocking up / down is determined by a CPU utilization threshold vs a time period. Llano's were factory set to only clock up if there was > 60% utilization for 500ms, and only clock down if there is <40% utilization for more then 2000ms. TB only activates if two or more cores are at their lower P states, I haven't had time to figure out exactly where the threshold was on Llano. I do know that if three cores are running at their P0 state then TB will never activate.Realizing the logic behind boosting you can see why we rarely see it activate on anything in Windows, even if it's a single threaded application. Windows is constantly moving it around which prevents the CPU from ever properly down clocking / boosting. Forcing processor affinity is a way to prevent windows from moving it around and thus allows for the conditions where TB can activate. I've done it multiple times on my A8-3550MX.[/citation]

I can confirm this. I was a fan of Simcity4, which is a single-threaded game that can drag a 6 GHz i7-3770K to its knee, even if all but one core was turned off.

My laptop has a quad core i7-720qm, which Windows 7 should've turned off two of the cores and allowed Turbo Boost to increase the speed of the heavily loaded core. Instead, it behaves like an idiot and constantly shifts SC4's thread from core to core, resulting in a consistent "25%" load on all of the cores, even when SC4's traffic simulator is bogged down due to its endless appetite for single-core processing power.

 

[lott11]

Forget it, you do not get it.
It has not thing to do with graphics.
It has to do with judgment.
It has to do being a journalist and being impartial.
The test dose not reflect all points of the testing.
It Just has to look good, to one side or the other.
A comment is one thing but inferring is another.
But never mind we have to be politically correct.

 

[blazorthon]

[citation][nom]lott11[/nom]Forget it, you do not get it.It has not thing to do with graphics.It has to do with judgment.It has to do being a journalist and being impartial.The test dose not reflect all points of the testing.It Just has to look good, to one side or the other.A comment is one thing but inferring is another.But never mind we have to be politically correct.[/citation]

You have yet to give any factual evidence to your claims... That is why I don't get it. So far, you've been plain wrong.