Is there any reason to use FX CPUs right now?

Page 7 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Feb 11, 2015
140
0
0
Actualy Haswell was a step back in efficency compared to IB, this allow the FX to compete with this gen.
LOL no no AMD FX is not competing with Intel Haswell LOL. My Core i5 4690K OC is a top dog and best price to performance gaming CPU. Meanwhile AMD FX is still trying to keep up with core i3 and getting pitied by Pentium Dual Core G3258 LOL .
 

Abwx

Lifer
Apr 2, 2011
11,885
4,873
136
LOL no no AMD FX is not competing with Intel Haswell LOL. My Core i5 4690K OC is a top dog and best price to performance gaming CPU. Meanwhile AMD FX is still trying to keep up with core i3 and getting pitied by Pentium Dual Core G3258 LOL .

In anything other than games your 4670K is not a very good opponent for a FX, and it is a less good one each month that pass, now if games are 99% of your usages that s another matter, be happy with your 4670K but just do not imply that your single usage, games, is a generality for perfs in other softs...

http://www.hardware.fr/focus/99/amd-fx-8370e-fx-8-coeurs-95-watts-test.html
 
Feb 11, 2015
140
0
0
Last edited:

ShintaiDK

Lifer
Apr 22, 2012
20,378
146
106
Actualy Haswell was a step back in efficency compared to IB, this allow the FX to compete with this gen, moreover given that softs get obviously more multithreaded, unfortunately AMD didnt get a smaller process that would had rendered relevant a Steamroller based FX, too bad since numbers suggets that Broadwell DT parts wont be more efficient than HW and hardly higher clocked.

Wrong, wrong and wrong.
 

Yuriman

Diamond Member
Jun 25, 2004
5,530
141
106
Actualy Haswell was a step back in efficency compared to IB, this allow the FX to compete with this gen, moreover given that softs get obviously more multithreaded, unfortunately AMD didnt get a smaller process that would had rendered relevant a Steamroller based FX, too bad since numbers suggets that Broadwell DT parts wont be more efficient than HW and hardly higher clocked.

I'm sorry, but this is untrue. Most of Haswell's gains only show up in the chips with hyperthreading, but it was a significant step forward in mobile in both performance and efficiency, and a fair step in desktops:

http://www.xbitlabs.com/articles/cpu/display/core-i3-4340-4330-4130_6.html#sect0

photoshop.png


truecrypt.png


x264.png


freemake.png


power-1.png


power-3.png



^ Doesn't look to me like a regression in Haswell i3's
 
Last edited:

Abwx

Lifer
Apr 2, 2011
11,885
4,873
136
No the 4690K will take any FX CPU for a run in pretty much any task. When OCed to 4.8Ghz the Intel 4690K cannot be touched by some FX crap that needs nuke power to operate. Even at mutitasking the Intel 4690K is much better than some AMD FX shiz ... http://www.bit-tech.net/hardware/2014/07/03/intel-core-i5-4690k-review/4

The FX can also be overclocked, and on average much higher than a i5, not a lot of i5 will do 4.8, i think that about none manage to do this frequency without liquid cooling, dont even dream about it on air, actualy you re using imaginary CPUs as comparative exemple, why not 6GHz while you re at it..?.

At equal frequencies an i5 is no match for a FX8 moreover when it s integer tasks, you would had noticed if you took attention to the link i provided and checked the integer softs results, and when overclocking the FX scale better than an i5.
 
Feb 11, 2015
140
0
0
The FX can also be overclocked, and on average much higher than a i5, not a lot of i5 will do 4.8, i think that about none manage to do this frequency without liquid cooling, dont even dream about it on air, actualy you re using imaginary CPUs as comparative exemple, why not 6GHz while you re at it..?..
The only thing imaginary is in thinking AMD FX is competitive with intel. My 4690K will do 4.8Ghz on higher end air cooling but on the FX you will need the higest end most exspensive air cooling or water to do 4.8 stable. Even then the 4690K @ 4.8 and the 8350 @ 4.8 the 4690K still pulls way ahead of the FX ship in the link I posted. In short you need an aggressively clocked AMD FX 8 core chip to even come somewhat close to a stock clock core i5 Devils Canyon in most tasks.
 

Sweepr

Diamond Member
May 12, 2006
5,148
1,143
136
No the 4690K will take any FX CPU for a run in pretty much any task. When OCed to 4.8Ghz the Intel 4690K cannot be touched by some FX crap that needs nuke power to operate. Even at mutitasking the Intel 4690K is much better than some AMD FX shiz ... http://www.bit-tech.net/hardware/2014/07/03/intel-core-i5-4690k-review/4

Don't even bother arguing, if the benchmark doesn't scale perfectly (or almost) with 8+ cores/threads then it's somehow considered Intel-biased by the usual suspects. It's easier to just pretend that per-core performance doesn't matter anymore like Hardware.fr does, might convince someone to buy the 220W turds.
 

Abwx

Lifer
Apr 2, 2011
11,885
4,873
136
I'm sorry, but this is untrue. Most of Haswell's gains only show up in the chips with hyperthreading, but it was a significant step forward in mobile in both performance and efficiency, and a fair step in desktops:

http://www.xbitlabs.com/articles/cpu/display/core-i3-4340-4330-4130_6.html#sect0

^ Doesn't look to me like a regression in Haswell i3's

I was speaking of perf/Watt, HW performs let say 10% better than IB on average but at a 20% higher comsumption on average, ie, perf/watt is 10% lower, the exemples you linked from Xbitlabs can be special case due to use of AVX2 in encoding tasks, in this latter case the perf/watt will be about the same but that s a corner case.

The result is that as much as perf/Watt is a concern the FX8 has less trouble competing against HW than against IB, even if the former increased absolute perf/MHz, also AMD get a little benefit from their 8C concept since softs are obviously better Mthreaded by the day, it s obvious on Hardware.fr softs suite that is about the same as two years ago but with more recent versions, the same 8T CPUs increased their lead over the 4T CPUs.
 
Feb 11, 2015
140
0
0
Don't even bother arguing, if the benchmark doesn't scale perfectly (or almost) with 8+ cores/threads then it's somehow considered Intel-biased by the usual suspects. It's easier to just pretend that per-core performance doesn't matter anymore like Hardware.fr does, might convince someone to buy the 220W turds.
ROFLMAO ! Yup pretty much ... If the game does not run well on an AMD FX rig then it's not the CPUs low IPC and per core performance to blame it's an unoptimized game engine LOL. I almost did buy the 220watt turn they call the FX 9590 cause i am an AMD fanboy but reality checked in and I had the common logic to understand that Intel is the way to go at leased until AMD comes out with something decent.
 

Abwx

Lifer
Apr 2, 2011
11,885
4,873
136
The only thing imaginary is in thinking AMD FX is competitive with intel. My 4690K will do 4.8Ghz on higher end air cooling but on the FX you will need the higest end most exspensive air cooling or water to do 4.8 stable. Even then the 4690K @ 4.8 and the 8350 @ 4.8 the 4690K still pulls way ahead of the FX ship in the link I posted. In short you need an aggressively clocked AMD FX 8 core chip to even come somewhat close to a stock clock core i5 Devils Canyon in most tasks.

You didnt check the link i provided and are talking uselessly, there s an overclocking section for said 8370E, wich is the opponent of the 4670k, get a look there :

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

Their ocking is on air and using Prime 95 to check stability, and that was an average sample compared to most sites.

Now about heat your 4670K will be much hotter with less watts since it has much lower die surface, as such it require an equally if not most efficent cooler, but i guess that you didnt knew about it, anyway it s a given that i5s overclock substancialy lower than FXs...
 
Feb 11, 2015
140
0
0
You didnt check the link i provided
Oh I did and it shows the usual ... AMD FX 9590 @ 5Ghz getting overtaken by a stock clocked Core i5 in every test except in the heat and power consumption tests in where the FX 9590 was the biggest loser.
 

Abwx

Lifer
Apr 2, 2011
11,885
4,873
136
It's easier to just pretend that per-core performance doesn't matter anymore like Hardware.fr does,

And yet you relied on their datas not so long ago since it suited one of your points, why suddenly is Hardware.fr irrelevant once their datas are at odd with your opinion..?.

Obviously you re not balanced at all, if only you would had argue why the sudden 180° turn, okay, but all i read is a blank statement substancied by a self contradicting argument, you ll concede that it doesnt help your credibility, if ever this word has a meaning for you...
 

Yuriman

Diamond Member
Jun 25, 2004
5,530
141
106
I was speaking of perf/Watt, HW performs let say 10% better than IB on average but at a 20% higher comsumption on average, ie, perf/watt is 10% lower, the exemples you linked from Xbitlabs can be special case due to use of AVX2 in encoding tasks, in this latter case the perf/watt will be about the same but that s a corner case.

The result is that as much as perf/Watt is a concern the FX8 has less trouble competing against HW than against IB, even if the former increased absolute perf/MHz, also AMD get a little benefit from their 8C concept since softs are obviously better Mthreaded by the day, it s obvious on Hardware.fr softs suite that is about the same as two years ago but with more recent versions, the same 8T CPUs increased their lead over the 4T CPUs.

Looking at performance per watt, I don't see a regression in the i3's:

power-1.png


power-3.png


x264.png


truecrypt.png


What are you basing this off of?
 

Abwx

Lifer
Apr 2, 2011
11,885
4,873
136
Oh I did and it shows the usual ... AMD FX 9590 @ 5Ghz getting overtaken by a stock clocked Core i5 in every test except in the heat and power consumption tests in where the FX 9590 was the biggest loser.

You are obviously lying and crapping/trolling the thread, here the link again and the average of the tests where the stock i5 does so well in your imagination, i let people check the individual results :

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


getgraphimg.php


Now you can go back trolling here and there but dont expect me to answer to what seems to be the worst bad faith one can imagine..
 

Abwx

Lifer
Apr 2, 2011
11,885
4,873
136
Looking at performance per watt, I don't see a regression in the i3's:
What are you basing this off of?

I was talking of the i7s since theses are the most competitive opponents of the FXs but the i5 should be worse than the i7s, dont know for the i3s since i didnt pay attention to thoses and that there are not as much datas as for the top parts.

http://www.hardware.fr/articles/913-9/cpu-consommation-overclocking.html

The same will be measured in softs like 7 Zip or any other integer softs that is not AVX2 optimised and boosted by 20% at least.

You can check at Computerbase the power delta in this review of IB and HW in CB and compute the perf/watt :

http://www.computerbase.de/2014-09/amd-fx-8370e-im-test/

Sigh the only troll here is AMDs FX silly excuse for a heater ... er I mean CPU.

Obviously it s the rage that the FX often trounce your i5 that motivate your posts, that s usual with people thinking that they bought the best of the best just to witness less expensive parts performing better, then the rage will go against the product that prove that you made a bad choice, typical, much easier to question AMD that to question your discutable choices..
 
Last edited:

MiddleOfTheRoad

Golden Member
Aug 6, 2014
1,123
5
0
The only thing imaginary is in thinking AMD FX is competitive with intel. My 4690K will do 4.8Ghz on higher end air cooling but on the FX you will need the higest end most exspensive air cooling or water to do 4.8 stable. Even then the 4690K @ 4.8 and the 8350 @ 4.8 the 4690K still pulls way ahead of the FX ship in the link I posted. In short you need an aggressively clocked AMD FX 8 core chip to even come somewhat close to a stock clock core i5 Devils Canyon in most tasks.

Any 8 core FX will bulldoze (pun intended) an i5 in scientific computing. It's a matter of 8 Threads versus 4. It's not even close. That is just a fact. Just look at the hardware used on World Community Grid. A Vishera core is only around an Ivy or Sandy in performance, but having twice the threads -- even the Haswell i5 struggles to generate the same level of points daily as an octacore FX. The king of the hill for desktop CPU's on BOINC is the i7 5960x -- because it can run 16 projects simultaneously. But an i5 is not a good chip for this usage, people need to spend the extra money for the i7 if they are doing this type of work. An FX-8320 or 8350 can run 8 projects at once, an i5 will only process 4.

And this is coming from someone who is running an i7 4790k...... But facts are facts.
 
Last edited:

Abwx

Lifer
Apr 2, 2011
11,885
4,873
136
Is the entire Tom's Hardware benchmark suite a special case too? Posting the same Hardware.fr results ad-aeternum won't make your false claims more credible. ;)

wh.png


http://www.tomshardware.com/reviews/core-i7-4770k-haswell-review,3521-18.html

Lol, you had to rely on almost blinded reviews, the plateforms are not the same and they measure power at the main, besides there s no idle power graph in this review, you think that it s by chance that they truncated the end of the curves comsumption.?..

On the other hand Hardware.fr measure CPU comsumption in isolation, this way they get rid of the plateform influence, but you didnt tell me why you were using Hardware.fr average scores the other day and are branding them irrelevant suddenly, what happened..?.

Check here at Computerbase.de, you have the idle power, load power under Cinebench as well as the Cinebech scores, you can compute perfs watt in this case, to spare you the hassle the difference is about 10%, in IB s favour of course.

http://www.computerbase.de/2014-09/...mm-leistungsaufnahme-volllast-cinebench-x-cpu

What s next, that CB is not relevant.?.Even for Intel CPUs.?.
 

SlowSpyder

Lifer
Jan 12, 2005
17,305
1,002
126
No FX is old, hot and overpriced when considering you need high end cooling and mobo to get the most out of them.


You're over generalizing. You can get a motherboard + FX + capable enough cooler to run the CPU at 4.4-4.5GHz for ~$200, (an i5 4670 is $225 on Newegg, alone). They aren't as efficient as Intel, certainly. But I don't think they're nearly as bad as some people make them out to be either. I think the power use and associated costs are not an issue for most people. If you run your CPU at 100% load 24x7, than I understand why that's important to you. If you're a typical gamer, the extra cost in power use is probably very little.
 
Last edited:
Feb 11, 2015
140
0
0
You're over generalizing. You can get a motherboard + FX + capable enough cooler to run the CPU at 4.4-4.5GHz for ~$200. They aren't as efficient as Intel, certainly. But I don't think they're nearly as bad as some people make them out to be either. I think the power use and associated costs are not an issue for most people. If you run your CPU at 100% load 24x7, than I understand why that's important to you. If you're a typical gamer, the extra cost in power use is probably very little.
I could have went with AMD FX 8350 over my Z97X 4690K setup but all I would have saved was $50 bucks so it was a no brainer to go with intel this time. If I were to go for the FX 9590 it would have cost $50 more than my Intel setup.
 

schmuckley

Platinum Member
Aug 18, 2011
2,335
1
0
What's the power consumption on a hex-core chip overclocked by 1.8Ghz?
Who cares?
Actually not that much..
It's a 95-watt chip..32nm.
Less than a Q6600 or FX 8-core..that's for sure.

Seriously..all you people crying about "energy efficiency" on CPUs and GPUs is ridiculous.
It's about the difference of 1 light bulb from lowest to highest power usage.
If AMD released a 200w chip tomorrow that outperformed Intel..I'd be all over it.