AMD CPU speculation... and expert conjecture

[juanrga]

That is what the slide #13 says.

^^ bolded the relevant part.

orly. the slide only shows some red bars and some x-axis fps value (but never how many fps each bar represents) and a few cpu and apu names. nothing else.

This is the trivial part of the message given by the slide, the interesting part is the rest, which I already commented.

And my reply was that usually one doesn't promote one product at the expense of another from the same company, unless the former product is going to replace/substitute the second.

there was no expense mentioned, and absolute zero mention of any replacement and/or substitution. i read the whole slideshow, all the way down to the disclaimer slide where they say that they might represent misinformation.

You must be taking this part of my post out of context and then misinterpreting it.

Measuring the bars and using the axis scale.

i call b.s. on this. show the exact calculations.

Calculation is trivial, FP = 50 + D*C, where C is the scale of the graph and D the size of the bar beyond the 50 mark. The values of both D and C vary from monitor to monitor (or from paper to paper, if you are printing the slide).

Evidently the cost is not obtained from slide #13. Spending $70 less for missing only a 8% of performance @ 1080p agrees with my concept of a 'fraction of the cost'.

8%? of what? is that even a real measurement? or did you grab that off of thin air? show the actual calculations.
your 'concept'(in reality, the lack of it and reality..) is baseless from the begining. the percentage number is wrong, and you are basing overall performance/price on a single (i.e. the narrowest sample it can be) data point which is a gpu bound offline single player benchmark where cpu performance is not even a factor And the igpu, the biggest draw, is rendered irrelevant due to the use of a powerful gaming discreet card.

8 % of performance. See bold part. The percentage is not measured, but computed. Computations given above.

In the first place, when I presented the slide by the first time. I said that the point was not to discuss if the slide was representative or not, because the fact that AMD had made that slide was much more important. We can discuss if the situation represented in that slide is the general case or only a special case, but this is unimportant. The important is that the slide reflects AMD intentions.

and amd's intention is to get oems interested in 6790k, which is new, late in the refesh cycle (i.e. a silicon dump), poor price/perf. that's why benchmarketing fluff was needed.
Nothing Else.

Sure that the intention was to get OEMs interested in the APU, but as said before you cannot do at the expense of other products, unless the other products are going to be abandoned.

Nobody makes a slide as that and present them to OEMs attending a talk if the goal is to sell lots of FX-6350 and 8350. The slide was clearly made with the goal of emphasizing the APU and deemphasizing the FX line. Now take your own opinion about that.

if you understand that amd's goal was to promote the apu then why are you even arguing? you got that part right. what you got utterly and absolutely wrong was that amd's goal was to sell 'lots of' fx6350 and fx8350 at the 6790k introduction event. this bit is self-explanatory.

Answered before twice.

The baseline is unimportant, because they are ES tests and the absolute scores don't agree with reviews of final silicon. Pay attention to the benchmarks of the CPU that I gave before. All them were run @1.8Ghz. Therefore you cannot take the absolute score obtained by Kaveri and compare to review articles of FX-8350 @4GHz. This makes no sense. What makes sense is if you have compare the score obtained by Kaveri to the score obtained by the ES Piledriver FX @ 1.8GHz. Then you obtain the percentages Kaveri was 30% faster than Piledriver clock for clock (if the leaked benchmark was legit).

did you read what you just wrote? "compare e.s. kaveri to e.s. pd" what the heck? i never compared the posted kaveri benchmark scores to retail fx8350 scores. comparing kaveri vs pd e.s. benchmarks is even less logical let alone at 1.8 ghz. i was discussing both the kaveri numbers and fx numbers from that bench.

Now you have lost me entirely.

Moreover, the bandwidth that eDRAM can give is unrelated to the effective bandwidth obtained in a real test that doesn't fit in the eDRAM. There is one reason why ordinary reviews of Iris Pro did only test low resolutions.

the edram in iris pro acts as an L4 cache, it's far from being unrelated.

You are right, poor wording from my part. It would say that the effective bandwidth is smaller that eDRAM bandwidth when the test doesn't fit on eDRAM.

The GP cryptographic test is coded in OpenCL and CUDA and run on the GPU. No AES no HSA. Same answer than above about absolute values doesn't matter.

you mentioned earlier that no info other than the chart was given out. how did you know that the gp crypto test is coded in opencl and cuda? can't opencl and cuda run aes algorithm? if hsa, the biggest feature of kaveri, was not even part of that bench - makes the bench irrelevent.

Because I know details about the GP cryptographic test. No, one does not code in CUDA or OpenCL to run AES extensions on the CPU. No the bench is not irrelevant because measures the compute performance of the GPU.

"Fraction of the cost" is a useless marketing term. You know, 1000000/1 is a fraction too. Product X is a fraction of the cost of product Y but product X cost 1000000 times more. It means absolutely nothing.

"Fraction of the cost" means a small part of the cost, i.e. cheaper.

http://www.thefreedictionary.com/fraction

http://dictionary.cambridge.org/dictionary/british/fraction

 

[gamerk316]

We don't know good mantle is, more specifically, how many more fps it gives over Nvidia ?

5 ? 10 ? 15 ? 20 ? 25 ? 30 ?

Probably a handful. The overhead realated to draw calls has been reduced significantly over the years, so in the grand scheme of things, we're talking *maybe* 1-2 FPS here. The time the GPU spends doing nothing due to the overhead of draw calls, in comparison to the 16ms screen display interval, is trivially small.

Point being: Optimizing draw calls is optimizing the wrong thing. You can improve performance 10x, but it won't be causing any significant effect to FPS by itself.

Probably between 30% and 50% more FPS. Therefore if the DX game gives 3 FPS the MANTLE version will give 1-2 FPS more ;-)

Yes, DX 11 reduced the overhead, but it is still about 5-10x and the reduction implies the use of batch calls. If you want to draw 100 independent objects the overhead returns.

Off topic somewhat:

http://www.tomshardware.com/reviews/geforce-gtx-780-ti-review-benchmarks,3663-18.html

Read this review yet?

Could this guy be anymore NVidia fanboy?

NVidia's GTX 780Ti allows the fans to spin at higher RPMs generating more noise, but that's the price you pay for more performance.

That's for a card that runs at 48 dB at full load! Yet, in the 290X review, the 55 dB at full load (a mere 7 dBs difference) was "intolerably loud, we could only run it in quiet mode"???

WTF! 55 dB is about the level of a normal conversation, 48 dB isn't much quieter than that at all...it's not like we're talking the difference between 48 and 70 dB or something absurd. There wasn't any coil whine or vibration from the 290X and it was "intolerably loud"?

That's completely BS!

We'll see what happens when the 290X GHz Toxic Edition comes...

/rant

To be fair, the new Catalyst driver ramped up fan speeds for the 290x, so they got even louder. Its quite clear AMD's reference fan isn't very good.

 

[gamerk316]

This has to be one of the biggest debate threads I have ever seen.

You weren't here for the BD rumor thread; we broke the forum a few times it got so big.

 

[de5_Roy]

This is the trivial part of the message given by the slide, the interesting part is the rest, which I already commented.

how is it trivial? you making up (i.e. flat out lying) facts is not trivial at all. please demonstrate, where in the slide #13, amd says that an a10 6790 provides >90% performance of an fx8350, at 'one fraction of cost'. unless there is some text hidden in the slide, you're talking b.s.
i asked before, ">90% of what? what tasks?" you avoided answering them. i also asked if there was an audio which might have discussed about that slide. you dodged that too. you provided no credible answers at all.
the single gaming bench in slide #13 does not represent an apu nor cpu task. in the bench, the gpu was doing majority the work. that's why using it a base for covering all cpu/apu workloads is a fallacy.

You must be taking this part of my post out of context and then misinterpreting it.

uh, no. if there is any context-twisting and misinterpretation, it is entirely on your part. since the begining.

Calculation is trivial, FP = 50 + D*C, where C is the scale of the graph and D the size of the bar beyond the 50 mark. The values of both D and C vary from monitor to monitor (or from paper to paper, if you are printing the slide).

again with calling vital information trivial.

ROFL. so if i am viewing the slide on my pc, or printing it, i am looking at varied performance results? [edited out snide remarks. too much room for sarcasm.]

anyway. what values did you use to calculate the exact fps values?

8 % of performance. See bold part. The percentage is not measured, but computed. Computations given above.

all you posted was an equation with 2 (i.e. two) unknown variables. how the heck do i solve a single equation with 2 unknowns? pardon my math skillls.
additionally, this 8% value, it lacks context. what kind of performance - overall or specific? the slide #13 cannot be representative of overall performance as i have explained multiple times already.

Sure that the intention was to get OEMs interested in the APU, but as said before you cannot do at the expense of other products, unless the other products are going to be abandoned.

the other cpus were not being undermined. and there was literally no mention of anything being abandoned.

Answered before twice.

you incorrectly answered. multiple times.

Now you have lost me entirely.

nope. if you can't answer, just say so.

You are right, poor wording from my part. It would say that the effective bandwidth is smaller that eDRAM bandwidth when the test doesn't fit on eDRAM.

my understanding was that the l4 cache would boost gpu compute performance, yet the benchmark shows the opposite, especially in trinity vs iris pro's case. chances are the kaveri sample (the one being benched) had much higher compute ability.

new amd price cuts:
http://www.cpu-world.com/news_2013/2013110701_AMD_cuts_prices_of_A6_and_A8_APUs_reveals_price_of_A10-6790K.html
fx9370 is at $224 now. $200 for the cpu, 24 bucks more for surefire o.c. clockrate as long as you run a supported mobo. i rank this one "much more palatable, may be get one if you're feeling adventurous" from previous "utterly worthless garbage". the rest of the centurion pile(geddit?) is still worthless until they all sell under $$180-200. as of the time of posting, newegg hasn't changed to the new price.

want moar of these kabini tablets, with 1080p screens:
http://www.techpowerup.com/194084/msi-unveils-the-w20-3m-amd-powered-tablet.html

 

[Embra]

I would have liked to have seen AMD put a little better reference cooler on the 290 & 290x, evn at the cost of a few more bucks. It would have given the cards a little more polish and quill the noise debate a little.

 

[juanrga]

This is the trivial part of the message given by the slide, the interesting part is the rest, which I already commented.

how is it trivial? you making up (i.e. flat out lying) facts is not trivial at all. please demonstrate, where in the slide #13, amd says that an a10 6790 provides >90% performance of an fx8350, at 'one fraction of cost'. unless there is some text hidden in the slide, you're talking b.s.
i asked before, ">90% of what? what tasks?" you avoided answering them. i also asked if there was an audio which might have discussed about that slide. you dodged that too. you provided no credible answers at all.
the single gaming bench in slide #13 does not represent an apu nor cpu task. in the bench, the gpu was doing majority the work. that's why using it a base for covering all cpu/apu workloads is a fallacy.

Your interpretation of the slide was a complete triviality. Nobody at AMD did that slide for that some OEM attending the talk was thinking "the slide only shows some red bars and some x-axis fps value (but never how many fps each bar represents) and a few cpu and apu names. nothing else". One doesn't make a slide to obtain that irrelevant reaction from an attendant. No sir.

Cannot you read a graph? Graphs are not made to give only numbers, for that task one can use a table with numbers. The slide #13 is saying in graphical form (no audio, no text...) that an A10 6790k provides >90% of the performance of an fx8350 for that gaming test. The "at one fraction of the cost" is not in that slide. I already explained to you, before, that it is a trivial conclusion from knowing the cost of each processor: the FX is more expensive.

If you really believe "the single gaming bench in slide #13 does not represent an apu nor cpu task" then try to repeat it using only the R9 280X and nothing more; no CPU no APU, only the Radeon

Of course, nobody said that that bench was representative of "all cpu/apu workloads". This is another flagrant misinterpretation of what I said.

Calculation is trivial, FP = 50 + D*C, where C is the scale of the graph and D the size of the bar beyond the 50 mark. The values of both D and C vary from monitor to monitor (or from paper to paper, if you are printing the slide).

again with calling vital information trivial.

ROFL. so if i am viewing the slide on my pc, or printing it, i am looking at varied performance results? [edited out snide remarks. too much room for sarcasm.]

anyway. what values did you use to calculate the exact fps values?

Reading a basic graph is not anything close to what I would call "vital", but a trivial task. Here in my country how to read a graph, how to extract a scale, graph interpolation, graph extrapolation... are simple tasks explained in the classroom at age of 12 years or so.

Who said you that the performance results vary if you see the slide in the monitor or in print. You are kidding true? Because I cannot believe that you cannot read a basic graph and still less believe that their meaning change with the reading medium (monitor vs print). Evidently the values of C and D vary, but one of them scales inversely to the other and thus the value FP is invariant, as must be.

I cannot say you the values, because I don't store this kind of irrelevant info. I wrote them in a note used them to obtain the FPs given in a previous message and then trash them after I did the computation.

8 % of performance. See bold part. The percentage is not measured, but computed. Computations given above.

all you posted was an equation with 2 (i.e. two) unknown variables. how the heck do i solve a single equation with 2 unknowns? pardon my math skillls.

They are not unknown, they are obtained from the graph. I said this to you before.

additionally, this 8% value, it lacks context. what kind of performance - overall or specific? the slide #13 cannot be representative of overall performance as i have explained multiple times already.

This was answered before. Also nobody said "overall".

Sure that the intention was to get OEMs interested in the APU, but as said before you cannot do at the expense of other products, unless the other products are going to be abandoned.

the other cpus were not being undermined. and there was literally no mention of anything being abandoned.

That is the message that the slide is saying. It is now evident for me that you don't see it. Of course AMD would not be saying during that talk, "we are abandoning..." no, specially when the successor has not been announced.

You are right, poor wording from my part. It would say that the effective bandwidth is smaller that eDRAM bandwidth when the test doesn't fit on eDRAM.

my understanding was that the l4 cache would boost gpu compute performance, yet the benchmark shows the opposite, especially in trinity vs iris pro's case. chances are the kaveri sample (the one being benched) had much higher compute ability.

No, the benchmark doesn't show the opposite. The results in that table are simply saying that AMD has a superior compute architecture than Intel, but this is something very well-known. The eDRAM in IrisPro cannot do miracles.

 

[Rum]

Yes, Db scale is not linear, it's exponential.

So 2 or even 3 Db do make a difference.

Anyway, I don't think the GTX780ti is a value winner, so I don't really think it will do fine. Specially with 3GB being a "premium class" video card. It will suck for big screens and multiscreens.

Cheers!

This link proves your point, the 290 is most definitely the best bang for the buck out of all the cards! http://forums.overclockers.co.uk/showthread.php?t=18555564

 

[8350rocks]

We don't know good mantle is, more specifically, how many more fps it gives over Nvidia ?

5 ? 10 ? 15 ? 20 ? 25 ? 30 ?

Probably a handful. The overhead realated to draw calls has been reduced significantly over the years, so in the grand scheme of things, we're talking *maybe* 1-2 FPS here. The time the GPU spends doing nothing due to the overhead of draw calls, in comparison to the 16ms screen display interval, is trivially small.

Point being: Optimizing draw calls is optimizing the wrong thing. You can improve performance 10x, but it won't be causing any significant effect to FPS by itself.

Probably between 30% and 50% more FPS. Therefore if the DX game gives 3 FPS the MANTLE version will give 1-2 FPS more ;-)

Yes, DX 11 reduced the overhead, but it is still about 5-10x and the reduction implies the use of batch calls. If you want to draw 100 independent objects the overhead returns.

Off topic somewhat:

http://www.tomshardware.com/reviews/geforce-gtx-780-ti-review-benchmarks,3663-18.html

Read this review yet?

Could this guy be anymore NVidia fanboy?

NVidia's GTX 780Ti allows the fans to spin at higher RPMs generating more noise, but that's the price you pay for more performance.

That's for a card that runs at 48 dB at full load! Yet, in the 290X review, the 55 dB at full load (a mere 7 dBs difference) was "intolerably loud, we could only run it in quiet mode"???

WTF! 55 dB is about the level of a normal conversation, 48 dB isn't much quieter than that at all...it's not like we're talking the difference between 48 and 70 dB or something absurd. There wasn't any coil whine or vibration from the 290X and it was "intolerably loud"?

That's completely BS!

We'll see what happens when the 290X GHz Toxic Edition comes...

/rant

To be fair, the new Catalyst driver ramped up fan speeds for the 290x, so they got even louder. Its quite clear AMD's reference fan isn't very good.

Well, it's good enough to cool the card, it's simply not as quiet.

Frankly...that's less a concern to me than many other things, I am very glad the drivers got the variance sorted out.

Everything said and done, I typically don't buy the reference design anyway. This last go around, when I bought the 7870XT, AMD didn't even have a reference design for it. Sapphire's 2 fan setup on the card looks remarkably like a Vapor-X setup (though it's not branded as such) and I have had no issues with noise or temps whatsoever...(it's even overclocked mildly to 1100 MHz)

EDIT: http://semiaccurate.com/2013/11/08/radeon-r9-290x-noisy-graphics-card/

S|A is reading my mind sometimes...I swear...

 

[gamerk316]

Yes, Db scale is not linear, it's exponential.

So 2 or even 3 Db do make a difference.

Anyway, I don't think the GTX780ti is a value winner, so I don't really think it will do fine. Specially with 3GB being a "premium class" video card. It will suck for big screens and multiscreens.

Cheers!

This link proves your point, the 290 is most definitely the best bang for the buck out of all the cards! http://forums.overclockers.co.uk/showthread.php?t=18555564

I'd argue the 770 GTX @ $329 is a better price/performance, with the 290 being the more powerful card of the two. Those would be my two top recommended cards right now in all honesty.

 

[8350rocks]

Yes, Db scale is not linear, it's exponential.

So 2 or even 3 Db do make a difference.

Anyway, I don't think the GTX780ti is a value winner, so I don't really think it will do fine. Specially with 3GB being a "premium class" video card. It will suck for big screens and multiscreens.

Cheers!

This link proves your point, the 290 is most definitely the best bang for the buck out of all the cards! http://forums.overclockers.co.uk/showthread.php?t=18555564

I'd argue the 770 GTX @ $329 is a better price/performance, with the 290 being the more powerful card of the two. Those would be my two top recommended cards right now in all honesty.

That honestly boils down to if you have the extra $70 to spend on a GPU, because if you can spend $400 on a GPU, and you don't buy the 290, you're crazy to buy anything else...

 

[de5_Roy]

Your reading of the slide was a complete triviality. Nobody at AMD did that slide for that some OEM attending the talk was thinking "the slide only shows some red bars and some x-axis fps value (but never how many fps each bar represents) and a few cpu and apu names. nothing else".

it's amusing that you think you can tell what someone else is thinking...
what i said was that the slide only represents an fps chart for a gpu bound game comparing existing amd products. and the chart really does not show exact values or anything else, making an educated guesstimate impossible.

Cannot you read a graph? Graphs are not made to give only numbers, for that task one can use a table with numbers.

i can, that's how i noticed the discrepancies.

The slide is saying in graphical form (no audio, no text...)

(excerpt from one of your initial posts)
ah, so there is no audio nor text. so what were you claiming before by -

In their talk they were saying to OEMs that the new APU gives >90% of the FX-8350 performance at one fraction of the cost. That is what the slide #13 says.

- was just a lie. how did you hear what amd reps told the oem reps from just a slide?

that an A10 6790k provides >90% of the performance of an fx8350 for that gaming test. The "at one fraction of the cost" is not in that slide. I already explained to you, before, that it is a trivial conclusion from knowing the cost of each processor: the FX is more expensive.

okay then, let's discuss the graphical form. the bolded part is the first time you precisely mention the gaming bench. earlier, every time, you were making blanket statements without specifying. still, the bars (rather intentionally) explicitly fail to deliver any concrete information about the precise amount of difference between the performances.
fx is more expensive - is another blanket statement. in reality, it depends on the sku being compared. fx6300 currently sells for $120 (newegg price), $2 less than a10 6790k's msrp of $122(retail price may differ). with fx6300 you don't have to pay for the igpu you won't use in a discreet card gaming pc, and you get more cores. your own comparison of using fx8350 is a form of benchmarketing lol. as for 6790k delivering >90%, show the exact calculation, with exact values, we'll talk then.

If you believe "the single gaming bench in slide #13 does not represent an apu nor cpu task" try to repeat it using only the R9 280X and nothing more; no CPU no APU, only the Radeon

who's taking things out of context now? (in case you misinterpret, it's always you.) why would i even do that? i never even slightly hinted at anything of that sort. poor straw man.

Of course, nobody said that that bench was representative of "all cpu/apu workloads". This is another flagrant misinterpretation of what I said.

answer the other part, why don't you? i.e. if it was not representative of all cpu/apu workloads that what exactly did it represent? you can't make blanket statements and then claim misinterpretation.

Calculation is trivial, FP = 50 + D*C, where C is the scale of the graph and D the size of the bar beyond the 50 mark. The values of both D and C vary from monitor to monitor (or from paper to paper, if you are printing the slide).

Reading a basic graph is not anything close to vital, but a trivial task. Here in my country it is explained in the classroom at age of 12 years or so.

i am so sorry. 😗

Who said you that the performance results vary if you see the slide in the monitor or in print. You are kidding true?

i kid you not. it's just that your own statement means exactly that.

Because I cannot believe that you cannot read a basic graph. Evidently the values of C and D vary, but one scale inversely to the other and thus the value FP is invariant.

riiight.. so what are those values?

I cannot say you the values, because I don't store this kind of irrelevant info. I wrote them in a note used them to obtain the FPs given in a previous message and then trash them.

ROFL. your whole argument is based on those. if you don't have them, then re-calculate please. do not try to avoid by calling them irrelevant. without the exact fps values, you cannot calculate the percentage that you keep claiming.

Their are not unknown, they are obtained from the graph. I said this to you before.

yea yea, so what are those "C" and "D" values? those are known to only you. i am not a mind reader and i've already apologized for my math skills. 😀

This was answered also. Nobody said overall.

no, this was not answered. you made blanket(and baseless) claims so i asked what they meant specifically.

That is the message that the slide is saying. It is now evident for me that you don't see it. Of course AMD would be saying during that talk, "we are abandoning... no when the successor has not been announced.

i don't see it because it's just not there. there is no mention of abandoning or phase out. how do you know from a slide what amd said? since, by your own account, there is no audio.
so you do understand that amd isn't talking about abandoning... yet you've been going on and on about exactly that.

[i am not replying to the(your) edited msg since i was too busy editing out sarcasm and sharp accusations off my own reply, and the edited post does not change anything from your original draft.]

 

[Rum]

Yes, Db scale is not linear, it's exponential.

So 2 or even 3 Db do make a difference.

Anyway, I don't think the GTX780ti is a value winner, so I don't really think it will do fine. Specially with 3GB being a "premium class" video card. It will suck for big screens and multiscreens.

Cheers!

This link proves your point, the 290 is most definitely the best bang for the buck out of all the cards! http://forums.overclockers.co.uk/showthread.php?t=18555564

I'd argue the 770 GTX @ $329 is a better price/performance, with the 290 being the more powerful card of the two. Those would be my two top recommended cards right now in all honesty.

The 770 is a good card no doubt but with the 290 only being 70$ more it takes the crown hands down even with that crappy cooler it has! Doesn't the cheaper 280X tie or beat the 770 in most reviews? All 3 of these cards are great offerings in the end it's just a matter of preference, with the biggest loser being the 780 non ti right now.

 

[de5_Roy]

edit: PS4 teardown video
http://www.wired.com/gamelife/2013/11/playstation4-teardown-video/

________

do you have a bulldozer cpu (or successor/derivative)???!!!??
do you have a scorpius(TM) gaming pc with all-amd parts??
no you don't. you don't have a compatible case.
no, amd doesn't endorse scorpius(TM) case (if they did i hope they'd pick one from aerocool 😛).
but wait. now you can have a bulldozer case!!!
http://www.techpowerup.com/194109/sharkoon-releases-the-bulldozer-midi-tower-pc-case.html
courtesy of sharkoon!
*at a distance, bd owners rejoice...*

another "bulldozer" case:
http://www.newegg.com/Product/Product.aspx?Item=N82E16811110017
sadly discontinued.

 

[jdwii]

Off topic but after reading toms latest article i would never buy a 290 or 290x until other coolers come out. I would say Amd needs to get their act together on cooling when it effects performance so heavily.

 

[juanrga]

what i said was that the slide only represents an fps chart for a gpu bound game comparing existing amd products. and the chart really does not show exact values or anything else, making an educated guesstimate impossible.

The graph shows exact values. I think you really mean it doesn't give the numerical labels for each bar, but they are not needed to make an "educated guesstimate". That is the power of a graph, it can convey info graphically. In fact, the graph transmit better the info than a table with the exact numbers :sarcasm:

The slide is saying in graphical form (no audio, no text...)

(excerpt from one of your initial posts)
ah, so there is no audio nor text. so what were you claiming before by -

In their talk they were saying to OEMs that the new APU gives >90% of the FX-8350 performance at one fraction of the cost. That is what the slide #13 says.

- was just a lie. how did you hear what amd reps told the oem reps from just a slide?

Bolded again the part that you don't pay enough attention. I was claiming that same than now. They said something very concrete using slide #13. Are you aware that info can be transmitted by other media beyond text or audio? A graph in a slide is a way to transmit info.

that an A10 6790k provides >90% of the performance of an fx8350 for that gaming test. The "at one fraction of the cost" is not in that slide. I already explained to you, before, that it is a trivial conclusion from knowing the cost of each processor: the FX is more expensive.

okay then, let's discuss the graphical form. the bolded part is the first time you precisely mention the gaming bench.

I believed that I had made my point clear enough since the first time that I commented in that slide. How wrong I was! :sarcasm:

earlier, every time, you were making blanket statements without specifying. still, the bars (rather intentionally) explicitly fail to deliver any concrete information about the precise amount of difference between the performances.

Nope, the graphic contains the exact info about the performance. What the graphic lacks are the labels, but the values can be obtained from 'reading' the graphic.

Who said you that the performance results vary if you see the slide in the monitor or in print. You are kidding true?

i kid you not. it's just that your own statement means exactly that.

I see now you are not kidding. Well, ok, then I will say in full form: you are plain wrong. I said that C and P vary, not that the result of multiplying both a summing the baseline (50) does. As a consequence, the performance is the same either if the graph is in your monitor or is in piece of paper.

Because I cannot believe that you cannot read a basic graph. Evidently the values of C and D vary, but one scale inversely to the other and thus the value FP is invariant.

riiight.. so what are those values?

What part of I got them from the slide #13 and then trash them because I was not going to use them anymore you don't understand?

I cannot say you the values, because I don't store this kind of irrelevant info. I wrote them in a note used them to obtain the FPs given in a previous message and then trash them.

ROFL. your whole argument is based on those. if you don't have them, then re-calculate please. do not try to avoid by calling them irrelevant. without the exact fps values, you cannot calculate the percentage that you keep claiming.

😆 The FPs values (I never said were exact values) were given before in one of my posts, three or four days ago.

And no I am not going to re-compute them. I will leave that to you as homework.

 

[juanrga]

Off topic but after reading toms latest article i would never buy a 290 or 290x until other coolers come out. I would say Amd needs to get their act together on cooling when it effects performance so heavily.

That means that the article aim worked for at least one person.

 

[lilcinw]

Off topic but after reading toms latest article i would never buy a 290 or 290x until other coolers come out. I would say Amd needs to get their act together on cooling when it effects performance so heavily.

That means that the article aim worked for at least one person.

And what was your interpretation of the aim of the article?

 

[blackkstar]

Off topic but after reading toms latest article i would never buy a 290 or 290x until other coolers come out. I would say Amd needs to get their act together on cooling when it effects performance so heavily.

It should be quite clear that Tom's is taking money for GPU reviews from Nvidia.

The press kit Nvidia gave review sites explicitly states that 290 series is not as consistent and it implies you should make a big deal out of it.

This whole thing comes from a marketing slide from Nvidia.

None of the other sites have been able to reproduce Tom's issue. In fact, other sites are finding that retail samples are actually faster than their review sample.

If no other sites find similar issues then it's safe to assume either Tom's bought a defective card or they screwed something up.

If Tom's did screw something up, no one will admit it. AMD won't admit it because then Tom's will tar and feather AMD every chance they get and Tom's won't admit it because it'd be a huge hit to their integrity.

Wonder if they'll swoop in and ban me for speaking like this, I got banned from [H]ardForum after having an account for YEARS and hundreds of posts because I questioned the almighty and all-knowing Brent and Kyle about using an overclocked, more expensive GTX 580 and then going "LOOK 7970 IS ONLY 15% FASTER ITS A DUD!"

Watch how Nvidia wrecks AMD in marketing and watch how they have so far. It'll make why AMD is so hesitant about an HEDT platform obvious.

Just this year alone, with AMD being very competitive with Nvidia, we've seen Nvidia:

1. Make a huge deal out of frame times right before 7990 was supposed to be released. 7990 would have beaten every single offering Nvidia had at the time. Instead it was a reject product no one wanted.

2. Offer OriginPC money and exclusive deals if they drop AMD and make a big deal about how AMD drivers are bad and the cards fail more than Nvidia. I thought Charlie was just talking crazy but OriginPC was the first to get their hands on new mobile parts

3. Now make a big deal about how GPU Boost lets reviewers get cherry picked cards and to call into question how GPU Boost works. This is coming from the company that botched GPU Boost in 320 drivers and made the drivers actually kill cards.

All of that in a single year, and that's just Nvidia. Intel is just as evil and will do things just as bad.

I don't think AMD can be competitive in the high end until they learn to play dirty. AMD should be telling reviewers to run benchmarks that call into question NVidia's memory bandwidth, image quality, and things like that. But they will just be the good guy and not play dirty and continue to get wrecked.

Intel will do the same if AMD releases a competitive part and it's already thought that it's already happening with Temash. For example sometimes you will find Temash models with a 2 cell battery while the Intel versions get the 4 cell battery, and they are the SAME LAPTOP MODEL, just different APU/motherboard.

 

[de5_Roy]

The graph shows exact values. I think you really mean it doesn't give the numerical labels for each bar, but they are not needed to make an "educated guesstimate". That is the power of a graph, it can convey info graphically. In fact, the graph transmit better the info than a table with the exact numbers :sarcasm:

at this point, you're just lying and trying to avoid providing any solid numbers.
yes, the graphs don't provide numerical values - i am amazed that it took you so long to realize that. only thing that chart shows is one graph longer/shorter than other ones without numerically stating why and by how much. you claim to calculate your percentage number without providing that vital information.

Bolded again the part that you don't pay enough attention. I was claiming that same than now. They said something very concrete using slide #13. Are you aware that info can be transmitted by other media beyond text or audio? A graph in a slide is a way to transmit info.

i paid enough attention to not take contents of that slide out of context. if you're interpreting differently, why not provide the basis of that information?

I believed that I had made my point clear enough since the first time that I commented in that slide. How wrong I was! :sarcasm:
....
Nope, the graphic contains the exact info about the performance. What the graphic lacks are the labels, but the values can be obtained from 'reading' the graphic.

since you claimed to obtain said values, why not provide them here? so far you have not provided anything credible to back up your claim. instead you resorted to straw man arguments and false claims.

I see now you are not kidding. Well, ok, then I will say in full form: you are plain wrong. I said that C and P vary, not that the result of multiplying both a summing the baseline (50) does. As a consequence, the performance is the same either if the graph is in your monitor or is in piece of paper.
.....
What part of I got them from the slide #13 and then trash them because I was not going to use them anymore you don't understand?

i'll reply to these at the same time: after multiple times of asking, you have repeatedly failed to provide any calculations, and solid numbers to back up your claims. now you tell me that you "'trashed' the results and you're not gonna use them anymore". that's just poor way of dodging.

you lied, got called out, failed to provide proof of your claims.

😆 The FPs values (I never said were exact values) were given before in one of my posts, three or four days ago.

And no I am not going to re-compute them. I will leave that to you as homework.

now you're saying that the fp/s values were given before without ever referencing them. AND you finally admit that those were not exact values - this completely contradicts everything you have presented until now. even in the begining of your latest post you claim of exact values.
if you'd just provided how you came to your conclusions with proper calculations and proof, i'd never have called you a Liar. but now i have to do it.

i cannot calculate it because there is nothing to calculate. the slide does not provide numerical measurements and the graph is impossible to use as a source. all i can see is that all 3 of the contenders delivered more than 50 fps but no way to calculate by how much. after a third look, the header says "battlefield 4: more performance with more cores" that pretty much pitches the fx6350 and fx8350- incorrect context and wrong bench (to push moar cores campaign) on amd's behalf. there is no way to verify the graph as correct since amd itself has disclaimed any responsibility.

edit2: with r9 290x, a stock athlon 750k delivers only 26.6 average fps, 22 fps min. in bf4 multiplayer benchmark (according to pclab.pl). meanwhile, a stock fx8350 delivers 39fps minimum - at 1080p resolution, with r9 290x. that's a relevant benchmark with numeric figure attached. people play bf4 for m.p. mode, not single player.

 

[kansasboy001]

in an A10 6790k legitimately does deliver 90% of the FX 8530's performance, thats great, but lets stop and think, it is STILL not as good as the 8350, where is my chip that gives 115% to 130% of the 8350? I don't care if it an apu. let's also think, the A10 6790k might give 90% in gaming ex, but in multi-threaded applications the 8350's 8 core are going to kill the A10's 4 cores.

 

[The Q6660 Inside]

in an A10 6790k legitimately does deliver 90% of the FX 8530's performance, thats great, but lets stop and think, it is STILL not as good as the 8350, where is my chip that gives 115% to 130% of the 8350? I don't care if it an apu. let's also think, the A10 6790k might give 90% in gaming ex, but in multi-threaded applications the 8350's 8 core are going to kill the A10's 4 cores.

^+1 Hell, even a 6350 will probably give a 6970K a roasting in multi-threaded apps.

 

[The Q6660 Inside]

in an A10 6790k legitimately does deliver 90% of the FX 8530's performance, thats great, but lets stop and think, it is STILL not as good as the 8350, where is my chip that gives 115% to 130% of the 8350? I don't care if it an apu. let's also think, the A10 6790k might give 90% in gaming ex, but in multi-threaded applications the 8350's 8 core are going to kill the A10's 4 cores.

the 6790k is a upcoming richland part, not kaveri. Kaveri is speculated to be 15-30%> than a 6800k.

Whoops 😗

 

[Cazalan]

in an A10 6790k legitimately does deliver 90% of the FX 8530's performance, thats great, but lets stop and think, it is STILL not as good as the 8350, where is my chip that gives 115% to 130% of the 8350? I don't care if it an apu. let's also think, the A10 6790k might give 90% in gaming ex, but in multi-threaded applications the 8350's 8 core are going to kill the A10's 4 cores.

^+1 Hell, even a 6350 will probably give a 6970K a roasting in multi-threaded apps.

Why would that be surprising?

$122 < $139 < $199

 

[The Q6660 Inside]

in an A10 6790k legitimately does deliver 90% of the FX 8530's performance, thats great, but lets stop and think, it is STILL not as good as the 8350, where is my chip that gives 115% to 130% of the 8350? I don't care if it an apu. let's also think, the A10 6790k might give 90% in gaming ex, but in multi-threaded applications the 8350's 8 core are going to kill the A10's 4 cores.

^+1 Hell, even a 6350 will probably give a 6970K a roasting in multi-threaded apps.

Why would that be surprising?

$122 < $139 < $199

I was assuming that the 6790K in the leaked Update slide was actually Kaveri, which it wasn't...

 

[GOM3RPLY3R]

Off topic but after reading toms latest article i would never buy a 290 or 290x until other coolers come out. I would say Amd needs to get their act together on cooling when it effects performance so heavily.

It should be quite clear that Tom's is taking money for GPU reviews from Nvidia.

The press kit Nvidia gave review sites explicitly states that 290 series is not as consistent and it implies you should make a big deal out of it.

This whole thing comes from a marketing slide from Nvidia.

None of the other sites have been able to reproduce Tom's issue. In fact, other sites are finding that retail samples are actually faster than their review sample.

If no other sites find similar issues then it's safe to assume either Tom's bought a defective card or they screwed something up.

If Tom's did screw something up, no one will admit it. AMD won't admit it because then Tom's will tar and feather AMD every chance they get and Tom's won't admit it because it'd be a huge hit to their integrity.

Wonder if they'll swoop in and ban me for speaking like this, I got banned from [H]ardForum after having an account for YEARS and hundreds of posts because I questioned the almighty and all-knowing Brent and Kyle about using an overclocked, more expensive GTX 580 and then going "LOOK 7970 IS ONLY 15% FASTER ITS A DUD!"

Watch how Nvidia wrecks AMD in marketing and watch how they have so far. It'll make why AMD is so hesitant about an HEDT platform obvious.

Just this year alone, with AMD being very competitive with Nvidia, we've seen Nvidia:

1. Make a huge deal out of frame times right before 7990 was supposed to be released. 7990 would have beaten every single offering Nvidia had at the time. Instead it was a reject product no one wanted.

2. Offer OriginPC money and exclusive deals if they drop AMD and make a big deal about how AMD drivers are bad and the cards fail more than Nvidia. I thought Charlie was just talking crazy but OriginPC was the first to get their hands on new mobile parts

3. Now make a big deal about how GPU Boost lets reviewers get cherry picked cards and to call into question how GPU Boost works. This is coming from the company that botched GPU Boost in 320 drivers and made the drivers actually kill cards.

All of that in a single year, and that's just Nvidia. Intel is just as evil and will do things just as bad.

I don't think AMD can be competitive in the high end until they learn to play dirty. AMD should be telling reviewers to run benchmarks that call into question NVidia's memory bandwidth, image quality, and things like that. But they will just be the good guy and not play dirty and continue to get wrecked.

Intel will do the same if AMD releases a competitive part and it's already thought that it's already happening with Temash. For example sometimes you will find Temash models with a 2 cell battery while the Intel versions get the 4 cell battery, and they are the SAME LAPTOP MODEL, just different APU/motherboard.

I'm going to tell you, personally from owning both types of cards and from seeing all of my friends computers, a majority of Nvidias cards are faster and easier to OC after the fact of what the press says than AMD.

I'm not going to say Toms is right, but their speculation seems pretty legit. The fact that the card runs over 90ºC with a super loud fan and only gets ~15% more performance than the cooler and quieter Titan blows my mind.

Then again, AMD has a history of heat resistance. I have an old 5570 that I roasted for two hours at ~110ºC with a worn fan the only runs 200-300RPM and is really loud, the card still runs as good as new. Unless AMD went 'dainty' on their latest GPUs, then no one should have a problem with the temperature.

The only thing is that when the Water Blocks are made for it and that can be applied, expect a large jump in price, and be known that the whole system and card to be cooled will probably cost more than the Titan itself, so then it's just a lose / lose situation.

Honesty with the ASUS GTX 770 DCuII at $340 right now, you'd get much better performance than the 290X if you put two of those 770s in SLI for ~the same if not a little more money than the other.

And from personal experience, I can tell you that at 3800 RPM with all my case fans turned off and my headset off, all I can hear from the card is a small breeze.

It's at 1260 (1058+202 Mhz) Core and 7500 (7010+490 Mhz) Memory clocks. Has yet to legitimately exceed 65ºC. ^_^

Just a recommendation. 😛

EDIT: As a side note, does anyone know how to get the 770 above 1212 voltage? It's really bothering me as the card has so much potential. I've heard the only way gets your card bricked -_-.