FX chips better than i5 for video editing?

[Abwx]

I wonder how many years will it take till an AMD CPU actually matches the ST performance of a 4.6GHz Pentium G3258 using 24/7 cooling.

And yet you didnt buy one apparently...because it is not what matters.

Looking at you sig you look content with about 3.46GHz Core 2 single thread perf, why suddenly doesnt it seems to be enough when it s others.?.

 

[Sweepr]

And yet you didnt buy one apparently...because it is not what matters.

Looking at you sig you look content with about 3.46GHz Core 2 single thread perf, why suddenly doesnt it seems to be enough when it s others.?.

My Yonah rig actually served me well back in 2006 (curently using Ivy Bridge and Haswell). Surprisingly, Core 2 Duo @ 3.46GHz's per core performance doesn't look bad at all compared to some of the most popular 2014 AMD CPUs.

 

[frozentundra123456]

On another thread a member has discovered that his G3258 at 4.6 has about half the FP throughput of a stock Kaveri, that is, the same as a 2C 4.0 Kaveri, i guess that no one will use IBT as single thread perf reference...



I would call this an automated reflex.




The vast majority use multithreaded softs, wake up , we are in 2014 with 2015 being at the corner...

Well, if the majority of users need this great multithreaded performance of the FX, why does it have what, maybe at most 5% of the total cpu market? I am sure you will think it is marketing, cheating, whatever by Intel, but perhaps most people dont live in this universe of yours where every task scales perfectly to 8 threads.

 

[Abwx]

My Yonah rig actually served me well back in 2006 (curently using Ivy Bridge and Haswell). Surprisingly, Core 2 Duo @ 3.46GHz's per core performance doesn't look bad at all compared to some of the most popular 2014 AMD CPUs.

And even Intel ones since a Haswell is only 10-12% better IPC wise in Fritz or 7zip than a core 2 but i guess that you thought that it was otherwise, hence your belief that it would do well only compared to AMD...

That s an appearance only since the system is somewhat cached in benchmarks, accounting for the I/O speed it is probably less smooth than even an AM1 plateform, but whatever, you are branding as weak CPUs that have consistently higher ST perfs.

Well, if the majority of users need this great multithreaded performance of the FX, why does it have what, maybe at most 5% of the total cpu market? I am sure you will think it is marketing, cheating, whatever by Intel, but perhaps most people dont live in this universe of yours where every task scales perfectly to 8 threads.

The FX accounted for about 10% of the desktop market in Q3 2013-Q1 2014, that s considerable for a single die and is a bigger part of the market than Intel s enthusiast plateforms, and this despite an unrelenteless viral marketing campaign by people who didnt even own the plateform and who where boasting mainly on power comsumption or irrelevant ST benches out of Mthreaded softs, the results in matter of power drain achieved by the 8370E were already possible for the vast majority of FX users.

As for MT softs , dont know, check Hardware.fr suite, all is not perfectly multithreaded but the results and averages are telling.

http://www.hardware.fr/focus/99/amd-fx-8370e-fx-8-coeurs-95-watts-test.html

 

[mrmt]

Well, if the majority of users need this great multithreaded performance of the FX, why does it have what, maybe at most 5% of the total cpu market? I am sure you will think it is marketing, cheating, whatever by Intel, but perhaps most people dont live in this universe of yours where every task scales perfectly to 8 threads.

FX is a failure even among people who actually NEED 8+ threads.

 

[SlowSpyder]

ST score is on line but MT score is well below the expectations, Computerbase.de scored 1711 at 4.7, you should be at about 1850.

http://www.computerbase.de/2014-09/amd-fx-8370e-im-test/2/#diagramm-pov-ray

Yeah, good enough to beat a 2700K in ST, and within 2% of what I predicted. FX CPUs do indeed seem to scale quite linearly on this bench.

You are absolutely right. I watched it more closely this time, some of the cores are going down to a 7x multiplier during the run (~1.5GHz). Looking into that now, score should be better...

Every time I think I have this motherboard figured out, it lets me know I still have more to learn. Anyway, tweaked my voltage, helped a bit, but still not right. Changed my LLC settings and we're golden. 1895PPS on eight cores at the same 5.106GHz (222x23).

 

[DominionSeraph]

but whatever, you are branding as weak CPUs that have consistently higher ST perfs.

irrelevant ST benches out of Mthreaded softs,

 

[Sweepr]

And even Intel ones since a Haswell is only 10-12% better IPC wise in Fritz or 7zip than a core 2 but i guess that you thought that it was otherwise, hence your belief that it would do well only compared to AMD...

Thanks god Haswell cores fare a lot better in other benchmarks compared to Conroe than those two you just handpicked.

Here's Conroe vs Sandy Bridge:

And here's Sandy Bridge vs Haswell: www.tomshardware.com/reviews/core-i7-4770k-haswell-performance,3461-5.html

As for MT softs , dont know, check Hardware.fr suite, all is not perfectly multithreaded but the results and averages are telling.

http://www.hardware.fr/focus/99/amd-...atts-test.html

Yup, they're telling that we might not see a Core i7 competitor so soon. It takes a 4.7-5GHz 220W to match the outdated 84W Core i7 4770K (overall) in the apps that it's supposed to shine, while Core i7 4790K is safe in the top spot and Haswell-E is miles ahead.

 

[Abwx]

Thanks god Haswell cores fare a lot better in other benchmarks compared to Conroe than those two you just handpicked.

Yup, they're telling that we might not see a Core i7 competitor so soon. It takes a 4.7-5GHz 220W to match the outdated Core i7 4770K, while Core i7 4790K is safe in the top spot and Haswell-E is miles ahead.

In cherry picking you are not that bad, what about the two other FXs since the 8350 account for about 70% at least of the 8C solds.?..

And i didnt know that the 4770K was outdated, you must live in an other universe, in these benches the 4770K is 9.5% ahead of the 8350 on average, not what i would call a massive advantage, we wont even talk of the price/perf, seems that it doesnt matter since, well, that s not a good metric...

 

[Sweepr]

In cherry picking you are not that bad, what about the two other FXs since the 8350 account for about 70% at least of the 8C solds.?..

Pretty valuable proposition as long as your usage scenario included a lot of apps that need 8+ threads, like most encoding/rendering apps tech websites usually test. If not, I'd stick with 4 faster cores.

And i didnt know that the 4770K was outdated, you must live in an other universe, in these benches the 4770K is 9.5% ahead of the 8350 on average, not what i would call a massive advantage,

It was replaced by a faster 4GHz Core i7 4790K (that some people called fake before launch) a few months ago, but I'm sure you might have missed that, that's why you're still talking about the Core i7 4770K in your comparisons. Also don't try to pretend that performance per core doesn't matter anymore, hint: that explains why Core i7 4790K delivers 40% faster gaming scores than the fastest competitor.

www.hardware.fr/focus/99/amd-fx-8370e-fx-8-coeurs-95-watts-test.html

 

[SlowSpyder]

Yup, they're telling that we might not see a Core i7 competitor so soon. It takes a 4.7-5GHz 220W to match the outdated 84W Core i7 4770K (overall) in the apps that it's supposed to shine, while Core i7 4790K is safe in the top spot and Haswell-E is miles ahead.

Since there is no perfect CPU, that being a CPU that scores infinity in benches while using zero watts and costs nothing, everything is a trade off. The i7 4770k is a great CPU, no doubt. Going by Newegg prices, it is also $335. You can get yourself an FX9370 for $205 with a $15 promo right now. That leaves you a lot of room for a quality cooler or bigger cooling to get those 5GHz+ clocks. Obviously that isn't for everyone, and one of the big trade offs of the FX is that it'll use more power (my Kill-o-Watt is showing about 250 watts of power use for my CPU while running these benches at 1.525v / 5.106GHz).

Like I said earlier, I'm not suggesting the FX is better, but I don't really think it is worse either, at least if you're not afraid to tinker a bit. There are just different trade offs to consider.

*edit - I used the 4770K since that was what was brought up earlier. The 4790K is in stock and only $5 more. Of course that'll be faster and improve bang for the buck ($5 more for ~14% higher clocks). But still, just different trade offs, none of them are perfect.

 

[Abwx]

Pretty valuable proposition as long as your usage scenario included a lot of apps that need 8+ threads, like most encoding/rendering apps tech websites usually test. If not, I'd stick with 4 faster cores.

It s not my habit to advise CPUs in tasks were they are not relevant, it s not the case for everybody here as some will advise their favourite brand without even looking at the context.

It was replaced by a faster 4GHz Core i7 4790K (that some people called fake before launch) a few months ago, but I'm sure you might have missed that, that's why you're still talking about the Core i7 4770K in your comparisons. Also don't try to pretend that performance per core doesn't matter anymore, hint: that explains why Core i7 4790K delivers 40% faster gaming scores than the fastest competitor.

www.hardware.fr/focus/99/amd-fx-8370e-fx-8-coeurs-95-watts-test.html

Both 4770K and 4790K are overpriced and have not as good price/perf ratio, moreover when accounting for the plateform, since SBridge Intel users had to go through a few MB updates while the AM3+ users just plugged the SKU with a bios revision at worst, even early chipsets like the 760/785G ended being certified and their owners can use the 8370E or any FX, set it to 8370E settings and get a very valuable upgrade while others are left spending scores of $ with one gen CPU MBs, and HW MBs doesnt make exception contrary to the infos that i did read at AT forums.

 

[Sweepr]

It s not my habit to advise CPUs in tasks were they are not relevant, it s not the case for everybody here as some will advise their favourite brand without even looking at the context.

Are you sure? You just picked two benchmarks where Haswell barely beats Conroe a few posts ago. Some might interpret that as a way to defend your favourite brand.

Both 4770K and 4790K are overpriced and have not as good price/perf ratio, moreover when accounting for the plateform,

You mean high-end CPUs with no direct competition might have worse perf/$ than cheaper CPUs from both Intel and AMD? Wow, you're so smart.

 

[Abwx]

Are you sure? You just picked two benchmarks where Haswell barely beats Conroe a few posts ago. Some might interpret that as a way to defend your favourite brand.

You mean high-end CPUs with no direct competition might have worse perf/$ than cheaper CPUs from both Intel and AMD? Wow, you're so smart.

You should put things in perspective, here when the FX Vishera was released : http://techreport.com/review/23750/amd-fx-8350-processor-reviewed/10

And when HW was released, same site, with a few doctored benches though when compared to their FX review :

http://techreport.com/review/24879/intel-core-i7-4770k-and-4950hq-haswell-processors-reviewed/12

Haswell didnt change the picture, the FX kept the same competitivity, and even better intrinsicaly since HW IPC improvements yielded negative returns compared to IBridge, HW is paradoxaly a lesser opponent than IB for the FX, unfortunately for AMD they are stuck in a half node process iteration, at 20nm 8C Steamroller based FX would be a quite different beast to deal with and it what was envisonned by AMD should had they kept property of their factories or benefited from a full node shrink.

 

[DominionSeraph]

It s not my habit to advise CPUs in tasks were they are not relevant, it s not the case for everybody here as some will advise their favourite brand without even looking at the context.

There's little need to look at the context when Intel will be the right choice for 99.999% of them.

 

[MiddleOfTheRoad]

FX is a failure even among people who actually NEED 8+ threads.

Yeah, tell that too all the Xeon owners that are getting owned by Bulldozers/Vishera on the World Community Grid.

The FX is still a great chip, but it only really shines on certain tasks.

 

[Phynaz]

Both 4770K and 4790K are overpriced and have not as good price/perf ratio

FX8350 is overpriced and have not as good price/perf ratio as Pentium G3258.

http://techreport.com/review/26735/overclocking-intel-pentium-g3258-anniversary-edition-processor/3

Next strawman?

 

[DominionSeraph]

Yeah, tell that too all the Xeon owners that are getting owned by Bulldozers/Vishera on the World Community Grid.

http://ark.intel.com/products/49942/Intel-Pentium-II-Xeon-Processor-400-MHz-1M-Cache-100-MHz-FSB

 

[Phynaz]

Yeah, tell that too all the Xeon owners that are getting owned by Bulldozers/Vishera on the World Community Grid.

Watcha talkin' about Willis?

http://boincstats.com/en/stats/15/host/breakdown/cpu/0/10/0

Perhaps you would like that sorted by credit per CPU second, where an FX8370 comes in 28th place.

http://boincstats.com/en/stats/15/host/breakdown/cpu/0/12/0

 

[Abwx]

FX8350 is overpriced and have not as good price/perf ratio as Pentium G3258.

http://techreport.com/review/26735/overclocking-intel-pentium-g3258-anniversary-edition-processor/3

Next strawman?

4.8GHz at 1.375V on 22nm process, you really find this serious, would you advise such settings.?.

Btw, at this frequency and using their numbers, that is 1.04V at 3.2Ghz and 30.8W measured with P95, they got this frequency at 80W TDP with P95 , quite an achievement in the waiting for the CPU to be damaged by this overvoltage, have you a few other exemples of the same barrel.?.

 

[AtenRa]

As for "encoding background video whilst playing games" (odd usage comparisons which only ever seem to spring up when comparing AMD quad's to Intel's duals... :sneaky it makes far more sense on ANY CPU to encode when you're not doing anything else (ie, when you're eating dinner, showering, taking the dog for a walk, browsing the web, or for huge workloads of dozens of queued up vids even leave it running overnight then set it to switch itself off when finished, etc). Common sense / intelligent time management is a good thing regardless of what CPU you own.

Yeap and by doing that you have more time your PC powered on using more energy and you spend more time to finish your work. But that is something reviews never show, they only concentrate on singly application performance.
Well sorry but we are in a age of multi-core CPU products, single core CPU era ended almost 10 years ago.:whiste:

 

[Erenhardt]

Not to mention people that are crying now because their g3258 can't clock above 4GHz within reasonable voltage.

Claiming 4770 as outdated in vishera vs haswell discussion is not the wise move.

I think his phenom is still a potent CPU. With some OCing it it can last a bit. Maybe when amd releases their new platform, the market will offer better options.

 

[mrmt]

Yeah, tell that too all the Xeon owners that are getting owned by Bulldozers/Vishera on the World Community Grid.
.

WCG is largely irrelevant when compared to the entire server and workstation market, you know, the market AMD lost after they started to sell FX.

 

[mrmt]

Since there is no perfect CPU, that being a CPU that scores infinity in benches while using zero watts and costs nothing, everything is a trade off. The i7 4770k is a great CPU, no doubt. Going by Newegg prices, it is also $335. You can get yourself an FX9370 for $205 with a $15 promo right now. That leaves you a lot of room for a quality cooler or bigger cooling to get those 5GHz+ clocks. Obviously that isn't for everyone, and one of the big trade offs of the FX is that it'll use more power (my Kill-o-Watt is showing about 250 watts of power use for my CPU while running these benches at 1.525v / 5.106GHz).

Like I said earlier, I'm not suggesting the FX is better, but I don't really think it is worse either, at least if you're not afraid to tinker a bit. There are just different trade offs to consider.

But these different trade offs is what makes a product better or not. From a commercial perspective Intel is hands down a better product. It gets you to a certain level of performance without tinkering, and with tinkering to levels of performance that the FX can't touch, all that with a smaller power consumption (and consequently a smaller BoM). It's simply a value proposition that AMD can't match.

That you are willing to overlook the FX shortcomings and put some of your hours in messing with your FX processor doesn't make any better, it's just capitalism proving itself again: You get what you pay for. And you are getting just that, an inferior product that you must invest your own time in order to make it a better than when it was sold for you.

 

[krumme]

WCG is largely irrelevant when compared to the entire server and workstation market, you know, the market AMD lost after they started to sell FX.

That already started when first gen core hit market. BD was not nessesary for the decline.
Add it took a few years before the k7 platform was ready for the server market. So AMD was only at the market for real for about 4-5 years or so.

The constantly comparing to that single short period of time (with soi / p4), is getting very old and is a bit of a strawman in this forum. Its beaten to death - lets move on.

AMD and Intel is not in any way equal and have, except for that short period, never been. Especially the years before where amd was just copying - and getting the k6 design from outside (k5 was just slow). AMD is getting way to much attention. The development is on Arm side with SS, qq and Apple. We should be discussing that - because its the future. 16 beefed up huge ARM cores for cheap and software to match. That should do it for video editing.