Skylake vs A10 AMD

[Teabaghead]

Thing is, with a 3.9 Quad core cpu with a great APU, why go Intel?

 

[Azuma Hazuki]

Because clock speed isn't everything, and Intel's i3 is going to run rings around anything AMD has aside from an FX83xx-series, for far less power usage too.

That said, the best bang for the buck is indeed an Athlon X4 and a GeForce 750Ti. You're looking at some fairly heavy cooling requirements, I believe 95W TDP for the processor and 50 or 75 for the GPU, but I've built thirstier and hotter systems in small cases.

Cooler Master's Elite 130 case is perfect for you http://www.amazon.com/Cooler-Master...=1448630751&sr=8-1&keywords=mini+itx+130+case I've built a computer with an i5 and a GTX 960 in one of these and it ran juuuust fine; if you're not using an optical drive the 110 will work too.

Pair this with an Earthwatts 380 green PSU and you're set

 

[Phynaz]

Thing is, with a 3.9 Quad core cpu with a great APU, why go Intel?

Because it's a downright slow cpu.

 

[DrMrLordX]

Thing is, with a 3.9 Quad core cpu with a great APU, why go Intel?

It's complicated, but the vast majority of software out there right now favors Intel's current design. The design of the chip you have now - the a8-7600 - is brilliant for some things, but not so wonderful for others. Someone else posted this link earlier:

http://www.anandtech.com/bench/product/1497?vs=1197

showing how the 7870k wins some benchmarks but loses some others. Sadly not many people benchmark the Kaveri/Kaveri refresh APUs, so they look at AMD's older FX processors and then try to make estimates accordingly, which is a poor idea.

I own an A10-7700k which the exact same chip as yours, just with unlocked multipliers and different base clockspeeds. You will probably enjoy that chip except in a few cases where things might get slow for you. If you want, start another thread and APU owners can give you some tips on how to get the most out of that APU, depending on how much trouble you want to go through tuning and testing that thing.

With very little effort, you can get it running @ 3.8 GHz in games with the iGPU fully loaded, which will make a difference in performance.

If the question were still up for discussion, I would back the decision to get the 7870k since it isn't THAT much slower than the 6100 (honestly), is sometimes faster than the 6100, and has superior iGPU so that you will be able to transition to games more comfortably. The only sticky issue with the 7870k is that you have to be prepared to go into the UEFI and alter the voltage settings in most cases. For whatever reason, motherboard OEMs want to overvolt the heck out of that chip. But that's easy to do.

 

[Abwx]

It's complicated, but the vast majority of software out there right now favors Intel's current design.

That s not even the case, Intel is favoured because most sites use outdated protocols that are typical of the single core era, curiously multitasking benches completely disappeared because the reviewers think that with several cores the applications throughput would scale accordingly while it s not true at all.

Since you have a Kaveri you can do the tests below and check if the throughput collapse like the Intel i3/i5s when two softs are running concurrently :

http://www.computerbase.de/2015-10/...gramm-multitasking-test-cinebench-plus-winrar

 

[AtenRa]

Because it's a downright slow cpu.

 

[monstercameron]

For normal people telling the difference subjectively might be tricky. And if you look at the benchmarks, you will save a few seconds here and there on Intel...so the Intel platform is better culmatively when you tabulate all the saved time. As for price, and igp prowess, the a10/a8 aren't bad or super unusable slow. It's just a slower CPU...

 

[Abwx]

And if you look at the benchmarks, you will save a few seconds here and there on Intel..

Intel is slower on both benches, so how can it be faster at the end of the day.?.

It would be interesting to launch the benches simultaneously with a X264 file that would require about the same time as povray in the individual tests, i m 100% sure that the i3 would be far behind.

 

[monstercameron]

No one buying these chips in this price segment really cares about heavy duty video editing or raytracers. They care about office apps, web browsers, multimedia and light gaming. Those benchmarks are kinda pointless to the everyman. But as I also said, subjectively it is difficult to tell them apart save for fps in games.

 

[AtenRa]

No one buying these chips in this price segment really cares about heavy duty video editing or raytracers. They care about office apps, web browsers, multimedia and light gaming. Those benchmarks are kinda pointless to the everyman. But as I also said, subjectively it is difficult to tell them apart save for fps in games.

 

[ShintaiDK]

Talk about cherry picking and your own numbers. The key word here is what, IGP accelerated?

Its funny if it was so good, why it sells so terrible bad. Let me give you a hint, its terrible, its slow and its rubbish.

And its easier just to post this:
http://anandtech.com/bench/product/1497?vs=1197

Going Intel pretty much means relatively twice as fast web browsing just as an example.

 

[Abwx]

Talk about cherry picking and your own numbers. The key word here is what, IGP accelerated?

Its funny if it was so good, why it sells so terrible bad. Let me give you a hint, its terrible, its slow and its rubbish.

And its easier just to post this:
http://anandtech.com/bench/product/1497?vs=1197

Going Intel pretty much means relatively twice as fast web browsing just as an example.

If it did sell as badly as you re fuding then Intel wouldnt have lost some marketshare in the DT segment last quarter..

As for browsing Aten Ra s tests above say that you are completly wrong..

Btw, cherry picking you said..?.

But that WebXprt is an Intel bench doesnt seem to be a cherry pick, how surprising, inst it..

Software and workloads used in performance tests may have been optimized for performance only on Intel microprocessors. Intel is a sponsor and member of the BenchmarkXPRT Development Community, and was the major developer of the XPRT family of benchmarks.

http://vrworld.com/2014/11/03/shades-sysmark-2001-intel-may-webxprt-problem/

When your product is crappy you can always rig the benches, wich is apparently a specialty at Intel.

 

[bystander36]

One thing is clear, the AMD equivalents are much slower for gaming and it's important to recognize what your purpose is for your purchase so you look at the right benchmarks.

 

[jhu]

Intel is slower on both benches, so how can it be faster at the end of the day.?.

It would be interesting to launch the benches simultaneously with a X264 file that would require about the same time as povray in the individual tests, i m 100% sure that the i3 would be far behind.

Right, because whenever I render something on my cluster I make sure to transcode on one of the nodes. I mean, who does this?

A more realistic demanding multitasking scenario is gaming+recording.

 

[bononos]

Right, because whenever I render something on my cluster I make sure to transcode on one of the nodes. I mean, who does this?

A more realistic demanding multitasking scenario is gaming+recording.

Yes he's been making outlandish statements lately like Intel is slower and even the i5's cannot handle more than 1 application when AMD has been on the losing end since Sandy/Ivy except for niche areas like slow gaming rigs/opencl apps using APUs. And the other guy is hawking his rigged benchmarks.
"do well in isolated benchmark but in real use these are poor performers"
" i3/i5s can eventualy handle a lot of threads only if a single software is used, the perf collapsing is due to loading the CPU with more than one application"

 

[frozentundra123456]

Yea, Intel is so far ahead in the benchmarks now that even cherry picking certain ones is not good enough for the ADF. Not they are reduced to trying to drum up artificial scenarios to discredit bench marking pretty much in its entirety. After all AMD "feels" faster, any objective data be damned. The benchmarks dont really count because applications X,Y,Z are not running while the benchmark is being taken. Makes sense to me, because I always try to load my cpu up as much as possible with other apps before running a cpu intensive task. /sarcasm

 

[Abwx]

Right, because whenever I render something on my cluster I make sure to transcode on one of the nodes. I mean, who does this?

A more realistic demanding multitasking scenario is gaming+recording.

Gaming + recording, with firewall, antivirus, wifi and a few tabs with flash animations.
That s closer to a real usage and certainly more ressource hungry during peaks than the tests i linked in some other threads.

Yea, Intel is so far ahead in the benchmarks now that even cherry picking certain ones is not good enough

LOL...

I guess that you close everything else once you start a game, that s a realistic way to use a PC, isnt it...

 

[mikk]

I do use Firewall, Antivirus and Wifi, there is basically zero CPU load all the time. I doubt this is a meaningful CPU load instance.

 

[Arachnotronic]

Gaming + recording, with firewall, antivirus, wifi and a few tabs with flash animations.
That s closer to a real usage and certainly more ressource hungry during peaks than the tests i linked in some other threads.



LOL...

I guess that you close everything else once you start a game, that s a realistic way to use a PC, isnt it...

You should buy an Intel Core i5 and actually use it. Serious question: have you ever owned a Core i5/i7 desktop CPU?

 

[frozentundra123456]

I do use Firewall, Antivirus and Wifi, there is basically zero CPU load all the time. I doubt this is a meaningful CPU load instance.

Exactly, I have all 3 running on a low end sandy i5 and my cpu usage hovers between 0 and 3 percent. Crap, you can run those on an atom tablet and still have minimal cpu utilization.

 

[Abwx]

You should buy an Intel Core i5 and actually use it. Serious question: have you ever owned a Core i5/i7 desktop CPU?

And likewise, did you ever owned a FX8xxx yourself.?.

Personaly i own neither an i7 nor a FX but i tested recently a friend s FX8350 wich is in heavy environment.

Since i m a piano player he asked me to test his huge synthesizers emulator and recording studio, wich i did while ignoring that his PC was rendering and encoding at the same time (with an X264 encoder and 3DS Max for the rendering), first time i ve seen something like 30GB on the task manager and yet it seemed like the CPU was mono tasking, FTR he own a i5 that he branded as not adequate.

As i pointed it the i7 is adequate for such heavy loading, so dont move the goalpost by including this CPU in the list that i criticised for their inability to be used this way.

Now you ll agree that CPUs below the i7 have some weakness that are not known by most people here, and wich i didnt think myself that this could be possible, like everybody i was assuming that tasks would be linearly scaling with the ressources.

Indeed it seems that there is a consensus even within the Intel supporters that the i7 is vastly superior to the i5 in real world even if on paper, and most benches, this latter has 80% of the the former s performance, otherwise there wouldnt be this kind of answers, including yours, to a member s question on the subject :

http://forums.anandtech.com/showthread.php?t=2455596

 

[Abwx]

Exactly, I have all 3 running on a low end sandy i5 and my cpu usage hovers between 0 and 3 percent. Crap, you can run those on an atom tablet and still have minimal cpu utilization.

Try downloading something with your Atom, we ll see if it s at a 3%...

Edit : Or eventualy transfering datas between two external drives...

 

[AtenRa]

IST to discredit benchmarks they are not Intel favored.

People they dont Multi-Task because they cant, if they could they would do it as well.

Multi-Tasking people do,

1: Gaming + Recording
2: Rendering + Gaming
3: Movie Editing (Finalizing) + Gaming
4: Video encoding + Gaming

Just because Core i3 and Core i5 users cannot do those workloads, doesnt mean anyone else doesnt do it as well.

And before you say anything, im not talking playing BF4 MP 64 player maps but something like playing a racing simulator or a turn-based game like Civilization etc.

 

[ShintaiDK]

Gaming + recording, with firewall, antivirus, wifi and a few tabs with flash animations.
That s closer to a real usage and certainly more ressource hungry during peaks than the tests i linked in some other threads.

You haven't supplied any evidence for this load usage. Most likely because you already know its so minimalistic that it can barely be counted.

I may only have 360mbit internet, but even downloading at full speed gives me pretty much 0 CPU usage over AC wifi with firewall on. Same goes for a bunch of edge tabs.

But again, I dont use an AMD CPU. So it may peak on those

 

[ShintaiDK]

IST to discredit benchmarks they are not Intel favored.

People they dont Multi-Task because they cant, if they could they would do it as well.

Multi-Tasking people do,

1: Gaming + Recording
2: Rendering + Gaming
3: Movie Editing (Finalizing) + Gaming
4: Video encoding + Gaming

Just because Core i3 and Core i5 users cannot do those workloads, doesnt mean anyone else doesnt do it as well.

And before you say anything, im not talking playing BF4 MP 64 player maps but something like playing a racing simulator or a turn-based game like Civilization etc.

People multitask all the time. You just have to setup some obscure situation to try and sell your AMD CPUs. Its funny you always get stuck in some sub 1% niche.