Haswell i3-4150 vs FX 8320/e for budget gaming rig?

[TheELF]

Give the 775 or whatever other quad a SSD and it will be snappier than a 4770K overclocked to 5GHz.

The Haswell is better at same drive, let say a HDD, due to the SATA interface mainly, certainly that memory play a little role but the HDD is the big bottleneck.

WOW.....ok.
So the speed of the processor does nothing for the speed of the computer,the last 35 or so years where all an intel sham with fixed benchmarks and we still have the same snappiness and speed as we had with the 8086.
We had no progress at all until we started to use 4 cores or more.
Ok dude.

 

[2is]

This snappiness phenomenon needs a meme created in its honor.

 

[TheELF]

This snappiness phenomenon needs a meme created in its honor.

Ohh.......snap.

 

[Erenhardt]

lol oh yeah? Literally hundreds of CPU reviews from dozens of sites and not a single one has numbers to illustrate the "snappiness" of AMD over Intel. Not only do they not have the numbers, they don't even mention the phenomenon.

This experience you're describing, doesn't exist outside your own head. That, by it's very definition, is you being delusional. Either that or you're simply making it up, which would make you a liar, so take your pic. Nice hole you've dug for yourself here.

Actually I remember a few mentions of that in reviews.
https://teksyndicate.com/videos/amd...s-3820-gaming-and-xsplit-streaming-benchmarks

https://teksyndicate.com/videos/amd-fx-8350-oc-vs-i5-3570k-oc-battle-continues

In both they mention more responsiveness from amd system.
I've seen it mentioned somewhere else, but I'm not sure where...(TH? not really sure)

 

[Abwx]

You said cinebench is real world and well... Intel wins, so...

Short memory, i said that it was adequate within a same brand, for cross comparisons in FP Povray is infinitely better.

Now are you actually done with me, or was that the delusions talking?

Perhaps that i have a lot of patience...

If you really want to see more benches, go ahead ant look up reviews right there at this very site. I guarantee you they're there even if I don't personally spoon feed them to you. Done digging yet? Oh, and they wont mention snappiness either.

I know enough decent sites but perhaps that i should mention them more often..

WOW.....ok.
So the speed of the processor does nothing for the speed of the computer,the last 35 or so years where all an intel sham with fixed benchmarks and we still have the same snappiness and speed as we had with the 8086.
We had no progress at all until we started to use 4 cores or more.
.

i quote myself for the purpose :

Give the 775 or whatever other quad a SSD and it will be snappier than a 4770K overclocked to 5GHz.

You know perfectly well that current CPUs are limited by their drive.

And no need of straws like the reference to the 8086, the MC68000 was much better...

This snappiness phenomenon needs a meme created in its honor.

Nice creation of yours, just dont pretend that i have something to do with it, i noticed that you and some other people are resorting to another tactic since you are rebuked easily when speaking technicaly, i guess that trying to stick monikers on targeted people is the habit of some professional viral marketers, it s a signature, so just keep on this route, it will be an aknowledgment.

 

[skipsneeky2]

Any older computer even going back to a pentium 4 will feel and run miles better with a ssd till you find out what the cpu can't handle. I was toying around with a p4 630 based Dell XPS400 with one of my ssds and well it runs incredibly fast till you pretty much bog the cpu doing something as simple as load a failblog page which could take up to 20 seconds before the lag disappears and you can scroll the page lol.

 

[PliotronX]

This snappiness phenomenon needs a meme created in its honor.

It was funny when it was thrown around a lot when the Rambus users refused to admit that RDRAM was overpriced for the Pentium 4 platform.

 

[2is]

Actually I remember a few mentions of that in reviews.
https://teksyndicate.com/videos/amd...s-3820-gaming-and-xsplit-streaming-benchmarks

https://teksyndicate.com/videos/amd-fx-8350-oc-vs-i5-3570k-oc-battle-continues

In both they mention more responsiveness from amd system.
I've seen it mentioned somewhere else, but I'm not sure where...(TH? not really sure)

My mistake, if you look at my previous post, I specifically mentioned professional review sites. I should have carried that word over in my subsequent posts.

 

[Headfoot]

Yes delusion, delusion when a brand need to pay a firm so that a bench give it 26% advantage from the start over the competition, delusion when one use a 2C 1.4 APU as argument to say that a 4C 2.05 APU will work no better.

Here the pathetic advantage that you re talking about, 100% enginered by Maxxon and their subsider :



That s the advantage, not due to uarch but simply to ICC and the rest of the tools they gave to Maxxon, surely for free.

That said i proposed to use POVRAY and keep CB only for a same brand comparison, is that what one would call biaised, given the numbers above.?.

This Cinebench cheats garbage has been repeatedly debunked. http://www.realworldtech.com/forum/?threadid=135978&curpostid=136051

The only "evidence" that ever existed for it was CPUID strings found in a hex editor after someone decompiled the binaries. Because having a string value to identify between AMD / Intel chips MUST BE cheating. It certainly couldn't be so that the program knows what name to display or any other number of reasons why you'd want to identify the user's processor in a processor benchmark program...

 

[Abwx]

This Cinebench cheats garbage has been repeatedly debunked. http://www.realworldtech.com/forum/?threadid=135978&curpostid=136051

The only "evidence" that ever existed for it was CPUID strings found in a hex editor after someone decompiled the binaries. Because having a string value to identify between AMD / Intel chips MUST BE cheating. It certainly couldn't be so that the program knows what name to display or any other number of reasons why you'd want to identify the user's processor in a processor benchmark program...

Brand it garbage as much as you want, numbers dont lie and dont need walls of text.

So that a bench favour a brand over another ones does not matter, i mentioned that Intel tool were used starting from CB 11.5 and gave a link that show the numbers, here they are , comparison of i3 4030u and A8
6410 with CB R10, 11.5 and R15 :


....................CB R10.....CB 11.5....CB R15

i3 4030U.......... 6285......2.03..........191

A8 6410........... 7010......2.04..........165

So with R10 the A8 is 11.5% (!) faster, come a new version, CB 11.5, that is Intel compiled, profiled and miraculously the i3 is now on par with the A8, but that s not good, Maxxon didnt work well enough for their mentor so they released R15 and now the i3 is 15% faster than the A8, but as you said it doesnt matter provided it favour the relevant brand, isnt it...

https://www.youtube.com/watch?v=PzLxCo5qofo

 

[TheELF]

Brand it garbage as much as you want, numbers dont lie and dont need walls of text.

Yeah numbers don't lie but that doesn't stop people from not understanding them.
Did you ever stop to think that way back in the days of R10 Hyperthreading was something completely new and that programs would not recognize the extra cores thus giving worse performance?
http://forums.cgsociety.org/showthread.php?t=565852
(post #5 he explains that the bike benchmark file gives reduced numbers)

And, oh no, what cheats they are, they brought out a new version that corrected this problem???No!!!!
And after that one that would work even better?

Man,some companies just hate us computer users.

 

[Abwx]

Yeah numbers don't lie but that doesn't stop people from not understanding them.
Did you ever stop to think that way back in the days of R10 Hyperthreading was something completely new and that programs would not recognize the extra cores thus giving worse performance?
http://forums.cgsociety.org/showthread.php?t=565852
(post #5 he explains that the bike benchmark file gives reduced numbers)

And, oh no, what cheats they are, they brought out a new version that corrected this problem???No!!!!
And after that one that would work even better?

Man,some companies just hate us computer users.

Dont try to deflect the subject, things are much simpler than that, Maxxon has a partnership with Intel, as shown in the video, probably that they were given the whole suite for free in exchange of advertising for Intel and it is clearly stated that among others :

https://software.intel.com/en-us/articles/optimization-notice#opt-en

Indeed the numbers above just show that it s the case, it s not by chance that i used two recent uarches that have all necessary instructions for scores comlparisons, you ll see the same pattern between Kabini and Baytrail with CB R15 reducing the difference compared to CB 11.5.

Now compare all thoses CPUs using Povray, wich is also a rendering task, and you ll have a quite different picture, yet povray is well optimised for Intel, just look at the difference between SB/IBridge and Haswell in this bench.

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

 

[Erenhardt]

If a single benchmark changes from considerably faster on one brand to considerably faster on other brand you know that there is something fishy about it.

Why the first version run slower on the second brand? Why it suddenly works faster? What changed? Why it changed? Which brand is actually faster?

 

[Headfoot]

Brand it garbage as much as you want, numbers dont lie and dont need walls of text.

https://www.youtube.com/watch?v=PzLxCo5qofo

Try reading.

You have 0 evidence, I have a quote from someone from Maxon. But "evidence" is useless in conspiracy theory debates so yall can keep believing whatever you want, as I'm sure you will do regardless

 

[Abwx]

Try reading.

You have 0 evidence, I have a quote from someone from Maxon. But "evidence" is useless in conspiracy theory debates so yall can keep believing whatever you want, as I'm sure you will do regardless

Maxxon guy did say that they optimise irrespective of CPU brand but what he doesnt say is that they have no control of the CPU dispatcher, this is implemented by Intel and they explicitely state it :

 

[Enigmoid]

If a single benchmark changes from considerably faster on one brand to considerably faster on other brand you know that there is something fishy about it.

Why the first version run slower on the second brand? Why it suddenly works faster? What changed? Why it changed? Which brand is actually faster?

It looks like R15 uses Haswell much better.

http://www.bit-tech.net/hardware/2014/08/29/intel-core-i7-5960x-review/8

CB 11.5 4770k vs. 3770k
8.12 vs. 7.91 -> 2.6% faster (AT's numbers: 11.8% are different from other sites for CB)

CB 15 4770k vs. 3770k
822 vs. 689 -> 19% faster

R15 posts stronger HW numbers than 11.5 relative to the field.

 

[2is]

Dont try to deflect the subject, things are much simpler than that, Maxxon has a partnership with Intel, as shown in the video, probably that they were given the whole suite for free in exchange of advertising for Intel and it is clearly stated that among others :

https://software.intel.com/en-us/articles/optimization-notice#opt-en

Indeed the numbers above just show that it s the case, it s not by chance that i used two recent uarches that have all necessary instructions for scores comlparisons, you ll see the same pattern between Kabini and Baytrail with CB R15 reducing the difference compared to CB 11.5.

Now compare all thoses CPUs using Povray, wich is also a rendering task, and you ll have a quite different picture, yet povray is well optimised for Intel, just look at the difference between SB/IBridge and Haswell in this bench.

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

He didn't deflect. He countered your conspiracy theories. Successfully.

 

[Abwx]

He didn't deflect. He countered your conspiracy theories. Successfully.

Successfully for whom is assuming that it s Maxxon that designed ICC.

The disclaimer above is explicit, Intel doesnt give the choice to customers that use their compiler, hence the disclaimer.

That said keep on interpreting things the fairy tales way.

 

[2is]

We've seen plenty examples of your delusions already, what you consider fairy tale is actually reality.

 

[TheELF]

Maxxon guy did say that they optimise irrespective of CPU brand but what he doesnt say is that they have no control of the CPU dispatcher, this is implemented by Intel and they explicitely state it :

Do you even understand what this picture says?
They only state the obvious, Intel can not guarantee that their compiler will produce code that runs the same everywhere.
And how could they guarantee anything else? Intel knows intel chips, they have the plans for them and everything,heck they designed them themselves,do they have the same knowledge of amd chips?No! So they cannot guarantee the same performance.

No connection between i3 4030U vs A8 6410 and AMD A10-6800K vs Intel Core i3-4330,totally different chips.

 

[ShintaiDK]

I cant believe he still post his nonsense.

Its getting worse than the SOI guy.

 

[Abwx]

Do you even understand what this picture says?
They only state the obvious, Intel can not guarantee that their compiler will produce code that runs the same everywhere.
And how could they guarantee anything else? Intel knows intel chips, they have the plans for them and everything,heck they designed them themselves,do they have the same knowledge of amd chips?No! So they cannot guarantee the same performance.

No connection between i3 4030U vs A8 6410 and AMD A10-6800K vs Intel Core i3-4330,totally different chips.

"Certain optimizations not specific to Intel microarchitecture are reserved for Intel microprecessors".


I think this is clear, it means that even some optimisation, ie , use of certains instructions, is possible with non Intel CPUs thoses optimisations will be only implemented for Intel CPUs.

Since Cinebench is ICC optimised and is encompassed by this disclaimer i suggested that it s relevant for a same brand CPUs comparison but not for comparison between different brands because Cinebench doesnt offer the guaranteed that it doesnt favour a brand over another one, dixit said disclaimer, period.

 

[ShintaiDK]

Since Cinebench is ICC optimised and is encompassed by this disclaimer i suggested that it s relevant for a same brand CPUs comparison but not for comparison between different brands because Cinebench doesnt offer the guaranteed that it doesnt favour a brand over another one, dixit said disclaimer, period.

So you completely ignore what the developer and have done and what compiler flags are used. Just to try and keep up with the conspiracy theory and keep on going that AMD got some performance they dont in your endless illusion. Same as the misguided statement year after year that FX CPUs will soon shrine due to their more cores. Just a matter of anytime soon, right? No, its not there and there is no conspiracy, move on.

 

[Abwx]

So you completely ignore what the developer and have done and what compiler flags are used. Just to try and keep up with the conspiracy theory and keep on going that AMD got some performance they dont in your endless illusion. Same as the misguided statement year after year that FX CPUs will soon shrine due to their more cores. Just a matter of anytime soon, right? No, its not there and there is no conspiracy, move on.

The develloppers have no control over the CPU dispatcher, even if they set the same path for all CPUs the CPU dispatcher will still create different paths according to the CPUID, this had been demonstrated by Guillaume Louel at Hardware.fr.

http://www.hardware.fr/articles/847-1/impact-compilateurs-architectures-cpu-x86-x64.html

http://www.hardware.fr/articles/847-16/conclusion.html

 

[ShintaiDK]

You forget the fastest AMD code comes from ICC on Windows.