AMD CPU speculation... and expert conjecture

[de5_Roy]

If you're going to base an entire supremacy argument on iGPU...then this discussion is an epic fail and you live in an alternate reality.

let's compare strictly cpu vs cpu - core i5 3570k/3770k has cpu, igpu, imc, pcie controllers all integrated. fx has.... cpu, imc...and nothing. i am not basing my claim on igpu, i am saying that the presence of it gives i5 3570k/i7 3770k an advantage fx8350 simply cannot beat in any way.
additionally, just because fx fails so badly does not make core i5/i7 'superior' (again, i never used that word). it might give that kind of illusion, but it's not real. calling this discussion names won't change that.

For benchmarks showing FX8350 with higher benches than i5-3570k please refer to my post with about 20 benchmarks a few pages back.

no need for that, i know very well how to skew benches to favor either brand. if you're refering to the xbitlabs, the resident amd fanboys denouced it not long ago. then again, none those are not strictly cpu benches, fx had the 'unfair' advantage of discreet gfx card's assist. fx has to have that, without it, it's useless.

Except that just about every AM3+ motherboard comes with built in video, even the $45 ones. No need to throw a discrete in just to see benches.

by amd3+ motherboards do you mean amd's motherboards with 970, 900x and 990fx chipsets? which ones of those have onboard gfx? as far as i know that some older, am3 (not am3+) mobos have on board gfx and many of them don't support fx. even if .. like, one of them supports fx as well as possesses on board gfx (i think it's msi 780g with radeon 3200 igp or something) - that's more like an outlier than norm, unlike sb/ib's igpu where almost all cpus have them and almost all chipsets (except p67, may be) support them.
yes, i have known all along about the possibility of fx being able to use on board gfx for display (on certain older motherboards) but should i have to spell it out for you (not you, cazalan) guys? that's pathetic. :lol:
from toms, bd reivew:

Enthusiasts still running previous-generation Thuban- or Deneb-based Phenom IIs may not want to adopt these FX-enabling BIOS updates. We hear from the motherboard vendors that the performance of an older processor could be negatively affected by extensive changes made to the AGESA.

FX processors will not work in an AM3-based board. AMD deliberately blocked that combination out in its BIOS. Now, what if you’re rocking an 890FX-based motherboard with a Socket AM3+ interface? AMD isn’t explicitly supporting such a configuration, though Socket AM3+/890FX boards do exist. Compatibility could be a roll of the dice, but there isn't any specific reason a vendor claiming interoperability can't achieve it.

it's a chance, at most.

 

[noob2222]

the problem is the way intel implemented the switch. If used to specify AVX for the maximum speed code, it will NOT RUN on any cpu that doesn't have AVX because you specified it not to write any lesser code.

Do you have any idea what are you talking about ?

Learn the difference between 'march' and 'mtune' .
Learn the difference between /Qx, /Qax , /arch:target,
Learn what is meant by code dispatcher. Learn what is meant by run time codepath selection.

Till you learn, stop talking out of your ass. You dont know the first thing about what you are so confidently blathering about.

you learn to read dickhead.

There is an option for setting the generic level higher or lower. For example, the options /arch:SSE3 /QaxSSE4.1,AVX will set the generic level to SSE3 and generate three versions of the code for the SSE3, SSE4.2 and AVX instruction sets. Non-Intel processors can only get the generic version, which will be SSE3 in this example. Code compiled with the /Qx option, for example /QxSSE4.1 will fail to run on non-Intel processors and processors without the specified instruction set.

http://www.agner.org/optimize/blog/read.php?i=49#73

Sorry, I missed the "non-Intel" and just said any cpu, you want to say im talking out of my ass, pucker up and kiss it.

 

[palladin9479]

Ok cease with the name calling and generally hostile behaviors. Lets keep this civil and somewhat professional please.

 

[sarinaide]

People need to chillax.

 

[jdwii]

Name calling such as that or "fanboy" usually comes up when they have no valid argument or important idea to bring to the table, then of course when one person upsets another they will end up getting mad at each other and usually take the name calling to the next level, but honestly its usually something we see on Youtube with the so called experts.

 

[noob2222]

no need for that, i know very well how to skew benches to favor either brand. if you're refering to the xbitlabs, the resident amd fanboys denouced it not long ago. then again, none those are not strictly cpu benches, fx had the 'unfair' advantage of discreet gfx card's assist. fx has to have that, without it, it's useless.

xbit is very biased. Prime example: "We measured the performance in Adobe Photoshop CS6 using our own benchmark"
Conclusion:
"AMD FX processors can’t boast high performance results in Adobe Photoshop. Without the resource-hungry filters, but during the typical versatile image processing eight- and six-core FX processors on Piledriver microarchitecture turn out even slower than the dual-core Core i3-3240. However, this can still be considered an achievement, because the previous generation FX processors got defeated even by Pentium G2120."

So ... what exactly did Xbit do to "their own benchmark" to move AMD from the top to being slower than a 3240?! That right there shows how relaible xbit is. Benchmarks designed specifically for the Intel cpus.

Not to mention their 8150 review took 3-4 monhts to find the right programs to show how bad it sucked.

As for calling fx useless without a graphics card ... some would say the same thing about Intel's IGP, usless for anything but web browsing.

 

[noob2222]

Name calling such as that or "fanboy" usually comes up when they have no valid argument or important idea to bring to the table, then of course when one person upsets another they will end up getting mad at each other and usually take the name calling to the next level, but honestly its usually something we see on Youtube with the so called experts.

ya, not sure what mayankleoboy's issue is with me, but seems like any time he gets a chance, he tries to drag out the worst side of me. If he wants to pretend its ok to throw out insults, I have no problem giving them right back. Normally it takes someone several tries, but his is an ongoing problem, well past just a few "gradeschool comments"

Ill try to refrain myself next time.

 

[sarinaide]

XBitLabs made me lol, two weeks ago I found how older benches on the exact same part different from the newer ones quite substantially in the amount of result boosting. We even ran our own benches to prove the number boosting to which we got a resounding no reply. Quite embarrassing when a "professional" online review site cannot get their own benches right.

This was in view of the A10 5800k and A8 5600K vs I3 3225 and I3 3220 article they had, we found on average they boosted the Intel results by around 12% to as high as 55% in some titles like F1 2012. We submitted F1 2012 results of our own to show the APU's gaming Ultra, 2x MSAA at 1080 in F1 2012 at 39 and 34FPS yet their results reflected it under low detail where they then somehow managed to get a i3 3225 to post 50FPS, which it doesn't achieve on ultra or lowest.

 

[de5_Roy]

Spoiler

xbit is very biased. Prime example: "We measured the performance in Adobe Photoshop CS6 using our own benchmark"
Conclusion:
"AMD FX processors can’t boast high performance results in Adobe Photoshop. Without the resource-hungry filters, but during the typical versatile image processing eight- and six-core FX processors on Piledriver microarchitecture turn out even slower than the dual-core Core i3-3240. However, this can still be considered an achievement, because the previous generation FX processors got defeated even by Pentium G2120."

So ... what exactly did Xbit do to "their own benchmark" to move AMD from the top to being slower than a 3240?! That right there shows how relaible xbit is. Benchmarks designed specifically for the Intel cpus.

Not to mention their 8150 review took 3-4 monhts to find the right programs to show how bad it sucked.

As for calling fx useless without a graphics card ... some would say the same thing about Intel's IGP, usless for anything but web browsing.

yup, that's what i was hinting at earlier.
behold how dual core pentium humiliates fx:
http://www.xbitlabs.com/images/cpu/fx-8350-8320-6300-4300/mathematica.png
looks like fx can not do math...
then,
http://www.xbitlabs.com/images/cpu/fx-8350-8320-6300-4300/photoshop.png

We measured the performance in Adobe Photoshop CS6 using our own benchmark made from Retouch Artists Photoshop Speed Test that has been creatively modified.

fear not! this is actually a good result. they say that earlier, pentium beat bulldozer fx, now just core i3 does it to piledriver fx, so it's okay!
then there is 'that' f1 bench comparing hd4k soundly beating radeon 7660d. ;-D
see, with a flick of the mouse, fx is bowing before pentium and core i3. is it believable, at this day?
what you really need to know is which software can really take advantage of your cpu's resources and you have a decent idea about performance.

agree about intel's igpu as well.
fx isn't useless without graphics card, it's just useless without any kind of discreet gfx, be it built on the mobo or a 7970 or anything in between. but if you yank out the discreet gfx off the intel test rig (with core i5 3570k/3770k), you can still see how the cpu bench did by connecting the display to the mobo's outputs.

 

[$hawn]

and ever since you have been trying to defend your 50% on average claim, even changing it from SB to all Intel cpus, trying to maintian the 50% lead.

Very untrue. I've mentioned that an AMD core was 66% of an SB core on average (note that I didn't use 'Intel was 50% faster', until JAYDEEJOHN pointed it out) in my very initial post. At the time, I was under the impression that most single threaded stuff would result in around the same value.
After you pointed out that it was only for iTunes AAC , and not on average, I apologized for not taking the average of a number of single threaded tests.
A few posts later, I again ran more calculations, and showed that for single threaded Cinebench, Intel was 37% faster. Again later, in reply to 8350rocks's post, I took the average of 3 single threaded tests, and proved that Intel averaged to be around 40% faster.
After that, I've even guessed AMD's average performance deficit against Intel to be in the order of 65% to 75%.

How does that amount to "defending my 50% on average claim" ? I only defended my claim for AAC encoding, but you don't wanna accept that. And yes, I'm very well aware that AAC != average, thank you.

all back-pedaling ... and a sudden "AAC is my favorite" defensive measure. I have never insulted you, just trying to offer advice that AAC != average performance.

"AAC is my favorite" was NOT sudden. I mentioned that to JAYDEEJOHN well before you started your argument with me, about 3 pages back.

Again, like I said, I acknowledged my '50% on average' mistake, and apologized. You call that back-pedaling? I think there's a huge difference between the 2 words.

so why all the

Why not post my whole sentence, here it is,
"I repeat again, all this is just my fancy guess work, and for all you know I could be very wrong. Do correct me if I am. My only intention was to put into perspective what a 5GHz centurion chip would look like against a SB-E chip."

Does this statement look like I am trying to defend my '50% on average' claim??

50% to 23% against SB is a huge margin of error, one you seem to continue to defend with changing the initial data.

You seem to want to cling to my 50% AAC value on purpose, and use it against me inspite my repeated attempts to make you understand. Your plan is easy to see.
Most of my calculations (not guestimations) are quite accurate. Please do point out any wrong ones, if you do find any. I'm not shy of learning from my mistakes.

Another funny thing is the choice in how the wording that is used. Intel being 51% faster is also the same as AMD only being 33% slower, Same as intel being 23% faster is only amd being 19% slower.

Hahaha, this is really a baseless accusation. AMD being 19% slower, or Intel being 23% faster IS MATHEMATICALLY THE SAME THING !! Or is this 'funny' fact beyond your ability to comprehend?

 

[JAYDEEJOHN]

Lets keep it civil
Show people, dont just criticize and complain

 

[palladin9479]

That made me think of the Laws of Power, specifically #9.

Law 9

Win through your Actions, Never through Argument

Any momentary triumph you think gained through argument is really a Pyrrhic victory: The resentment and ill will you stir up is stronger and lasts longer than any momentary change of opinion. It is much more powerful to get others to agree with you through your actions, without saying a word. Demonstrate, do not explicate.

Law 10

http://www2.tech.purdue.edu/cgt/courses/cgt411/covey/48_laws_of_power.htm

Some downright scary sh!t, more so when you realize their right.

 

[$hawn]

you learn to read dickhead.

http://www.agner.org/optimize/blog/read.php?i=49#73

Sorry, I missed the "non-Intel" and just said any cpu, you want to say im talking out of my ass, pucker up and kiss it.

MODERATORS, especially Brett928S2

So how come this guy walks free after posting something like this, while I got banned once upon a time? Are rules only for specific people?

 

[mayankleoboy1]

ya, not sure what mayankleoboy's issue is with me, but seems like any time he gets a chance, he tries to drag out the worst side of me. If he wants to pretend its ok to throw out insults, I have no problem giving them right back. Normally it takes someone several tries, but his is an ongoing problem, well past just a few "gradeschool comments"

Ill try to refrain myself next time.

The issue is that you have only a vague idea of how compiler optimizations and switches work. But still you shout at the top of your voice about things you have no idea about.
Repeatedly showing the same ignorance is bound to be called out in this forum.

 

[sarinaide]

Wrong thread. I want THG of old

 

[lilcinw]

MODERATORS, especially Brett928S2

So how come this guy walks free after posting something like this, while I got banned once upon a time? Are rules only for specific people?

Brett is retired if I am not mistaken.

Personally I think those abusing the lack of filter could use a vacation to cool off but that is not my decision.

 

[anxiousinfusion]

On a topic-related note I believe today is the day we find if WCCFTech was correct in their April 18-19 Richland launch. But prepare to be disappointed.

 

[sarinaide]

On a topic-related note I believe today is the day we find if WCCFTech was correct in their April 18-19 Richland launch. But prepare to be disappointed.

Yeah you will be, the release is only in June

 

[Cazalan]

yes, i have known all along about the possibility of fx being able to use on board gfx for display (on certain older motherboards) but should i have to spell it out for you (not you, cazalan) guys? that's pathetic. :lol:

I don't have any dice in the benchmark argument. I just went to newegg and searched for "motherboard AM3+" and the first page all had on-board video.

That's what struck me as odd because a decade ago building network booting F@H linux farms it was just as easy to find cheap boards with integrated video. ATI Radeon 3000 was a popular one.

 

[$hawn]

MODERATORS, especially Brett928S2

So how come this guy walks free after posting something like this, while I got banned once upon a time? Are rules only for specific people?

Brett is retired if I am not mistaken.

Personally I think those abusing the lack of filter could use a vacation to cool off but that is not my decision.

Retired??

 

[$hawn]

Yeah you will be, the release is only in June

June?!! Oh crap!!
Any idea when richland lappies might be available in the Indian market dude? I was under the impression that I could grab one by that time, but now I really doubt it. September looks more realistic

 

[Cazalan]

Yeah you will be, the release is only in June

When AMD said Richland was shipping in January I was expecting to see them out sooner. Not good to launch right on top of Haswell when there's going to be so much press/marketing for that.

 

[de5_Roy]

I don't have any dice in the benchmark argument. I just went to newegg and searched for "motherboard AM3+" and the first page all had on-board video.

none of the am3+ specific chipsets e.g. 970, 990fx have onboard gfx. but, some am3+ motherboards with older chipsets support fx. and some of those support on board video. afaik, none of these are officially supported by amd (from toms bd review). this one below is one of them:
http://www.asus.com/support/CPU/1/24/M5A78LM_LX_PLUS/#support_Support_CPU_FX-8350%28FD8350FRW8KHK,4.0GHz,8C,125W,rev.C0,AM3+%29
another chipset was 880g iirc.

 

[lilcinw]

@$hawn
That is the impression he gave me in this thread.

 

[sarinaide]

Yeah you will be, the release is only in June

June?!! Oh crap!!
Any idea when richland lappies might be available in the Indian market dude? I was under the impression that I could grab one by that time, but now I really doubt it. September looks more realistic

I was under the impression that April 18 was the date, but it appears the notebooks and ultrabooks will be out a month prior to Richland.

Yeah you will be, the release is only in June

When AMD said Richland was shipping in January I was expecting to see them out sooner. Not good to launch right on top of Haswell when there's going to be so much press/marketing for that.

I am really wanting it to be wrong but it seems like all dates are pointing towards a May/June release. Today the 7990 is released iirc.