AMD CPU speculation... and expert conjecture

[juanrga]

This is the ordinary version of the x264 software in Haswell

The difference between the hardware is minimal. Now what happens when a new version of the software optimized for the new instructions in Haswell is used?

The review (Extremetech) claims that the the drop-off in Pass 2 is due to the newness of the executable "we’re betting it’s a minor issue that’ll resolve with a tweaked version. Regardless, the R2334 binary delivers far more performance than the older program."

AMD has presented HSA-accelerated x265 at APU13:

I am amazed by the double standard used in this thread. When Intel develops video software optimized for its own hardware, that is fine. However, when AMD develops video software optimized for its own hardware, then that is the poor thing in the world and AMD is "only a software company". LOL!

 

[szatkus]

Szatkus, what games? Those that appear in the well-known "optimized for Intel" list that Intel publicities in its own site? Those that always appear benchmarked in sites known by their bias?

None of those games were optimized for Intel.

They are.

No, they're not!!!11

Ok, I'm not wasting my time on that type of disscussion. You have better imagination than I.

 

[gamerk316]

http://www.headline-benchmark.com/results/87080cac-179d-4a2c-a2a9-cb1e90667a3f

CPU detected right, RAM isn't.

 

[Ags1]

Win! Thanks gamerk 🙂. I do not yet get the RAM from registry, its stored in some kind of weird array. But its not a priority - the RAM is not used to index the results, while CPU name is, so I'm now getting the essential information. On the subject of names, AMD makes my life hell with the vague GPU name they give in OpenGL. For example, R9 cards will say they are "R9 200 Series" - nothing to say if they are 280/280X/290/290X. Drives me mad, but I don't have a solution yet. Nvidia give proper names, btw.

 

[gamerk316]

Nope, If I want play games at 1080p at medium settings testing CPUs at low settings and 720p is misleading. Moreover, as commented before, the tendency in games will be towards offloading the CPU generating GPU-bounded games.

Now you talk two different things. You can either:

A: Bench the CPU at the settings where the difference in CPU is the greatest

OR

B: Bench the CPU at settings where everyone actually plays at.

The problem with the second, of course, is any differences in CPU power is suppressed due to the GPU bottleneck in games at high settings. Running at lower settings shows the difference in power between two CPUs when said bottleneck is removed. Frankly, telling me CPU A and CPU B perform the same in game X with GPU Y is meaningless; what if I have a faster GPU, do the two CPU's still perform the same? Woops, don't have a clue. By testing at minimal settings though, you see which one is truly more powerful, even if the FPS numbers are meaningless.

Frankly, both ways should be done; min settings to show the order of CPU's in terms of power on said game, and at max settings to determine typical FPS with a given hardware setup.

 

[gamerk316]

Also

http://www.headline-benchmark.com/cpu/4923915864375296/5074136640520192

Lost some performance, but I'm assuming that's due to the new version of the SW. Will play some more tonight and try to break it 😛.

EDIT

On second review, the CPU *almost* detected right:

Intel R Core TM i7 2600K CPU 3 40GHz (8)

Versus

Intel(R) Core(TM) i7-2600K CPU @ 3.40GHz (8)

Lost the Parans and @ symbol. Not sure that will mess up the indexing or not, but you might be able to clean that up easily enough internal to the program. No idea why it isn't getting detected right the first time though...[Maybe run as Admin? Will try later tonight, need to run to work now]

 

[Ags1]

Yuka had some success wearing a black hat with the last version. That's why I went to an EXE wrapper and set process priority to high. Good luck!

Regarding the result variance, it's hard to say what the margin of error is for any given system.

^^^^ I can map names to a preferred normalized name on the server, but killing the special chars might be a good idea. Also for packing the the names into URLs.

 

[Reynod]

...Am I the only one here getting a strong sense of deja-vu?

The [strike]Bulldozer[/strike] Kaveri rumor thread, with Gamerk2 playing himself, noob2222 playing himself, and juanrga playing the star role of JF-AMD.

I thought so myself too ... however, please note that not once did JF-AMD EVER misrepresent AMD benchmarks ... he cherry picked and often talked a lot about power envelope under demanding server loads ... but not once did he misrepresent AMD.

He is a friend of mine and John Fruehe now works elswhere ... as the VP of NEXT IO after AMD, and currently working as a consultant in the server industry and running a blog ... and riding his mountain bike wherever there is a hill handy. I imagine he is getting to spend more time with the family as well ... after more than 5 years jumping in and out of planes.

You may pick on his lack of hair but do not call him a liar.

He left this site because we did not control the trolls well enough at the time.

Notice the lack of trolls lately ???

 

[juanrga]

^^Intel's success doesn't depend on hsa being successful. All of intel's software could fail and they would still make 4B per year.

AMD on the other hand is pushing lower end hardware in favor of HSA. Thats the problem. If AMD software fails, they have only a low end processor and a mid level gpu.

It is untrue that AMD kaveri is "only a low end processor". Your claims about Intel are also wrong. Intel Xeon cannot compete with IBM Power CPUs in the HPC market and the Xeon Phi can only barely compete with Nvidia GPUs using AVX extensions (Intel software instruction for his own heterogeneous approach). They are losing money in the HPC market and pushing hard their own heterogeneous approach to get that market.

At the same time, Intel is pushing hard its own heterogeneous approach (based in AVX extensions) in both desktop and mobile: Haswell, Broadwell, Skylake... To compete with AMD and HSA hardware/software and to compete with ARM (who is also a founder member of HSA). Intel is giving market to both, specially to ARM.

The notice of Google considering switching to ARM technology increased the stock of ARM whereas the Intel stock dropped in the same day. If Google makes the switch Intel will be in serious trouble.

You are painting an all-rose picture for Intel and an all-black picture for AMD. You are exaggerating and clearly on purpose.

 

[juanrga]

http://wccftech.com/amd-working-hynix-development-highbandwidth-3d-stacked-memory/
amd apu is on roll.. they know what they need on how to do it right.

Excellent finding! As I said before AMD is working in an ultra-high-performance APU with stacked RAM to compete with similar designs by Nvidia and Intel. Interesting that AMD choses HBM, whereas both Intel and Nvidia chose HMC. In principle HMC is faster,* but HBM is developed by JEDEC

And now a leaked slide

* The table in the wcctech link only reports HMC 1.0 bandwidth, HMC 2.0 is 320GB/s

 

[Reynod]

But the future for AMD is bleak ifthis gamble doesn't pay off ... but Intel have kazillions in the bank.

From what I can see Kaveri is not destined to be a monster of a CPU ... rather it is targeted at a particular "niche" i nthe market AMD feel will make them money, appeal to a particular range of devices.

It doesn't pay to get too blue, red or green Juan ... they are just soul sucking corporations who want your money.

 

[Reynod]

Intel is the biggest soul sucker of them all, with proven cases aof anti-trust in a number of countries ... so I try to avoid spending any of my hard earned cash on them.

Surprise surprise AND and NVidia are no better ... both have produced dodgy, hot GPU's in the past which didn't work properly and they both have misrepresented benchmarks, and gotten caught.

The trick in this sordid life is to get the best bang for your buck when you can.

When you die you can't take your PC with you either.

I have a mate who plans to be buried with his boat and 4WD ... but thats a story for another day.

 

[gamerk316]

Broadwell-EP news, why not?

http://www.pcper.com/news/General-Tech/Intel-Broadwell-EP-Xeon-E5-v4-Because-Why-Not

 

[juanrga]

But the future for AMD is bleak ifthis gamble doesn't pay off ... but Intel have kazillions in the bank.

From what I can see Kaveri is not destined to be a monster of a CPU ... rather it is targeted at a particular "niche" i nthe market AMD feel will make them money, appeal to a particular range of devices.

It doesn't pay to get too blue, red or green Juan ... they are just soul sucking corporations who want your money.

The Kaveri CPU is not destined to be an all-replacement for an i7-4770k CPU, but neither it can be labeled as "low end CPU 37% slower than Bulldozer" as some here pretend. The truth is in the middle.

I am not of any team! In fact I would like to make my own personalized team taking the best from each chipmaker: my ideal system would be an ultra-high-performance APU with some CPU parts from doubling an i5-4670k (i.e. 8 cores), the HSAIL from Kaveri CPU, a GPU mixture of the GCN and CUDA architecture, with Intel open drivers level support and documentation, hUMA, stacked RAM from HMC, a modified ARMv8 or modified AMD64 ISA,... In fact, I plan to write an article about it some day just for fun 😍

 

[de5_Roy]

What is AMD’s PowerTune 2.0 and what does it do?
http://semiaccurate.com/2013/12/16/amds-powertune-2-0/
has some good bits about how power gating works and limits to chip construction.

i can't recall the last time i saw an amd reference card look so... cute (regardless of performance, temps and acoustics)
http://techreport.com/news/25780/109-radeon-r7-260-graphics-card-coming-next-month
priced very well, compared to 7770's launch price. the reason i'm posting this because i suspect bonaire might have xdma (in reality, i want it to have xdma. that, and the capability to act as a physics accelerator). what if it can cfx with kaveri a10?!?!?! except.. that cfx bridge connector dims my hopes.

de5_Roy, you started to post insults and lies after it was shown that you are unable to perform a 10-year old calculation. I have ignored your posts lately until that you launched another ad hominem against me one page ago.

heh. what lies? please point those out like i pointed out your lies, along with proof for thinking so. i can't help if you feel insulted by that.

...like your ad hominem attack right in the post above? my posts don't contain anything like that. if you think they do, please point those out with explanation. i'll correct myself.
oh..do you mean that comparison with jf-amd? last page's.. that was not ad hominem, simple fact. i read jf-amd's posts and many of those were far, far more informative (some were quite educational and helpful to me) than anything you have ever posted. ever. times infinity. comparing a former amd executive to someone like you is an insult to the former executive and a huge compliment to you. i posted because mention of jf-amd.

and uh.. you didn't ignore my posts. all of your replies had baits in them, that's why i didn't reply back.

the rest of my post was simple observation (apart from the kaveri bit) on troll/shill/c.a.l.f. behavior. that happens around every new launch. it's like a periodical event. iirc, last time some guy asked everyone to buy amd shares right before they lost more than half their value.

 

[-Fran-]

But the future for AMD is bleak ifthis gamble doesn't pay off ... but Intel have kazillions in the bank.

From what I can see Kaveri is not destined to be a monster of a CPU ... rather it is targeted at a particular "niche" i nthe market AMD feel will make them money, appeal to a particular range of devices.

It doesn't pay to get too blue, red or green Juan ... they are just soul sucking corporations who want your money.

The Kaveri CPU is not destined to be an all-replacement for an i7-4770k CPU, but neither it can be labeled as "low end CPU 37% slower than Bulldozer" as some here pretend. The truth is in the middle.

I am not of any team! In fact I would like to make my own personalized team taking the best from each chipmaker: my ideal system would be an ultra-high-performance APU with some CPU parts from doubling an i5-4670k (i.e. 8 cores), the HSAIL from Kaveri CPU, a GPU mixture of the GCN and CUDA architecture, with Intel open drivers level support and documentation, hUMA, stacked RAM from HMC, a modified ARMv8 or modified AMD64 ISA,... In fact, I plan to write an article about it some day just for fun 😍

To add a bit here, I am looking for the cheapest Notebook I can find, but with most features it can have. Ironically, there's no AMD near the Celerons at the lower end of the spectrum.

The E2 line and the A4 line is more expensive than Intel's with the same features for low end. Thinking the notebook won't be used for gaming, it is hard to pass on the higher price (ironically) of the notebooks. In particular, the new 22nm line form Intel is looking amazing in the low end. Check the Pentium 2117U and the Celerons in that range.

Anyway, point in case, AMD needs Kaveri for the mid market ASAP. From a quick look, they're mostly tied everywhere, so marketing makes the difference. And we all know who wins by default on that one.

Cheers!

 

[Krnt]

Yeah as always the stupid OEMs aren't placing enough amd based notebooks or netbooks, don't even talk about the good ones, so their prices rise so high that make them look like a niche product.

Also the good processors for that type of products are almost nonexistant, since they are nowhere to find on the market even if they have been released almost a year ago.

 

[juanrga]

Nope, If I want play games at 1080p at medium settings testing CPUs at low settings and 720p is misleading. Moreover, as commented before, the tendency in games will be towards offloading the CPU generating GPU-bounded games.

Now you talk two different things. You can either:

A: Bench the CPU at the settings where the difference in CPU is the greatest

OR

B: Bench the CPU at settings where everyone actually plays at.

The problem with the second, of course, is any differences in CPU power is suppressed due to the GPU bottleneck in games at high settings. Running at lower settings shows the difference in power between two CPUs when said bottleneck is removed. Frankly, telling me CPU A and CPU B perform the same in game X with GPU Y is meaningless; what if I have a faster GPU, do the two CPU's still perform the same? Woops, don't have a clue. By testing at minimal settings though, you see which one is truly more powerful, even if the FPS numbers are meaningless.

Frankly, both ways should be done; min settings to show the order of CPU's in terms of power on said game, and at max settings to determine typical FPS with a given hardware setup.

Double review is a good idea, and some sites do that, finding that A10-6880k is ~40% behind i5-2500k at low settings and 720p/1024p, but only ~8% behind at high settings and 1080p. All this using a 7970 GHz Ed with Windows 7 and 'older' games. The differences are smaller with modern games under W8.1 or SteamOS.

 

[juanrga]

I am not of any team! In fact I would like to make my own personalized team taking the best from each chipmaker: my ideal system would be an ultra-high-performance APU with some CPU parts from doubling an i5-4670k (i.e. 8 cores), the HSAIL from Kaveri CPU, a GPU mixture of the GCN and CUDA architecture, with Intel open drivers level support and documentation, hUMA, stacked RAM from HMC, a modified ARMv8 or modified AMD64 ISA,... In fact, I plan to write an article about it some day just for fun 😍

To add a bit here, I am looking for the cheapest Notebook I can find, but with most features it can have. Ironically, there's no AMD near the Celerons at the lower end of the spectrum.

The E2 line and the A4 line is more expensive than Intel's with the same features for low end. Thinking the notebook won't be used for gaming, it is hard to pass on the higher price (ironically) of the notebooks. In particular, the new 22nm line form Intel is looking amazing in the low end. Check the Pentium 2117U and the Celerons in that range.

Anyway, point in case, AMD needs Kaveri for the mid market ASAP. From a quick look, they're mostly tied everywhere, so marketing makes the difference. And we all know who wins by default on that one.

Cheers!

Same problem here 🙁 only can find E1 or A4 models and the prices are not attractive.

 

[gamerk316]

Nope, If I want play games at 1080p at medium settings testing CPUs at low settings and 720p is misleading. Moreover, as commented before, the tendency in games will be towards offloading the CPU generating GPU-bounded games.

Now you talk two different things. You can either:

A: Bench the CPU at the settings where the difference in CPU is the greatest

OR

B: Bench the CPU at settings where everyone actually plays at.

The problem with the second, of course, is any differences in CPU power is suppressed due to the GPU bottleneck in games at high settings. Running at lower settings shows the difference in power between two CPUs when said bottleneck is removed. Frankly, telling me CPU A and CPU B perform the same in game X with GPU Y is meaningless; what if I have a faster GPU, do the two CPU's still perform the same? Woops, don't have a clue. By testing at minimal settings though, you see which one is truly more powerful, even if the FPS numbers are meaningless.

Frankly, both ways should be done; min settings to show the order of CPU's in terms of power on said game, and at max settings to determine typical FPS with a given hardware setup.

Double review is a good idea, and some sites do that, finding that A10-6880k is ~40% behind i5-2500k at low settings and 720p/1024p, but only ~8% behind at high settings and 1080p. All this using a 7970 GHz Ed with Windows 7 and 'older' games. The differences are smaller with modern games under W8.1 or SteamOS.

Which is a somewhat typical result. GPU bottleneck at high settings hiding the CPU difference, which is at least 40% for that given game (possibly more, if you could get settings lower via GPU driver tweaks). Run at low settings where CPU power is FAR more important, then the differences in power (and power reserve) becomes clear.

Which infers that unless AMD increases performance by 40%, Kaveri won't be catching the 2500k, at least in that particular game.

So understand what review sites are benching: They typically bench how a given processor(s) perform at given settings in a given benchmark. They do NOT directly compare the relative processing power of said processors when doing these benchmarks. This leads to false impressions, like a 6880k being only 8% slower then a 2500k.

 

[jdwii]

This is the ordinary version of the x264 software in Haswell

The difference between the hardware is minimal. Now what happens when a new version of the software optimized for the new instructions in Haswell is used?

The review (Extremetech) claims that the the drop-off in Pass 2 is due to the newness of the executable "we’re betting it’s a minor issue that’ll resolve with a tweaked version. Regardless, the R2334 binary delivers far more performance than the older program."

AMD has presented HSA-accelerated x265 at APU13:

I am amazed by the double standard used in this thread. When Intel develops video software optimized for its own hardware, that is fine. However, when AMD develops video software optimized for its own hardware, then that is the poor thing in the world and AMD is "only a software company". LOL!

This looks pretty cool can't wait to see that, should also fix some problems with slower internet speeds and HD videos. We are due anyways we've been using the same old x264 for so long.

 

[juanrga]

Broadwell-EP news, why not?

http://www.pcper.com/news/General-Tech/Intel-Broadwell-EP-Xeon-E5-v4-Because-Why-Not

This part caught my attention:

They expect close to 1 TFLOP per socket of double precision FP performance. This meets or exceeds the performance available by Kaveri including its GPU. Sure, the AMD solution will be available over a year earlier and cost a fraction of the multi-thousand-dollar server processor, but it is somewhat ridiculous to think that a CPU has the theoretical performance available to software render the equivalent of Battlefield 4's medium settings without a GPU (if the software was written with said rendering engine, which it is not... of course).

Kaveri has 856GFLOP of single precision performance. Double precision is not disclosed by AMD but expect less than 428GFLOP. The part about rendering Battlefield 4 must be a joke or something.

 

[juanrga]

Nope, If I want play games at 1080p at medium settings testing CPUs at low settings and 720p is misleading. Moreover, as commented before, the tendency in games will be towards offloading the CPU generating GPU-bounded games.

Now you talk two different things. You can either:

A: Bench the CPU at the settings where the difference in CPU is the greatest

OR

B: Bench the CPU at settings where everyone actually plays at.

The problem with the second, of course, is any differences in CPU power is suppressed due to the GPU bottleneck in games at high settings. Running at lower settings shows the difference in power between two CPUs when said bottleneck is removed. Frankly, telling me CPU A and CPU B perform the same in game X with GPU Y is meaningless; what if I have a faster GPU, do the two CPU's still perform the same? Woops, don't have a clue. By testing at minimal settings though, you see which one is truly more powerful, even if the FPS numbers are meaningless.

Frankly, both ways should be done; min settings to show the order of CPU's in terms of power on said game, and at max settings to determine typical FPS with a given hardware setup.

Double review is a good idea, and some sites do that, finding that A10-6880k is ~40% behind i5-2500k at low settings and 720p/1024p, but only ~8% behind at high settings and 1080p. All this using a 7970 GHz Ed with Windows 7 and 'older' games. The differences are smaller with modern games under W8.1 or SteamOS.

Which is a somewhat typical result. GPU bottleneck at high settings hiding the CPU difference, which is at least 40% for that given game (possibly more, if you could get settings lower via GPU driver tweaks). Run at low settings where CPU power is FAR more important, then the differences in power (and power reserve) becomes clear.

Which infers that unless AMD increases performance by 40%, Kaveri won't be catching the 2500k, at least in that particular game.

So understand what review sites are benching: They typically bench how a given processor(s) perform at given settings in a given benchmark. They do NOT directly compare the relative processing power of said processors when doing these benchmarks. This leads to false impressions, like a 6880k being only 8% slower then a 2500k.

First, I wrote "games", plural. The above percentages are averages.

Second, this comparison was made under Windows 7 (which is optimized for Intel). AMD APUs are 5-15% faster under W8.1 or SteamOS that under W7, thanks to a new scheduler than understand the CMT architecture of AMD modules, closing the artificial gap with Intel chips.

Third, the comparison used 1600MHz RAM for both, when Richland stock memory speed is 2133MHz. Add another 5% by using stock memory timings for Richland.

Therefore, AMD increasing hardware performance by ~ 20% is enough to caught the i5 in so-called CPU benchmarks. The rest 10-20% comes from using stock memory and an operative system not optimized for Intel 😉.

Fourth, the result of the 6880k being only 8% slower than a 2500k at 1080p at high settings is real and measured. If you play those games at that resolution and settings, that is the performance than you obtain. It is not some imagined score.

Fifth, as explained again and again and again and again in this thread. Future games (i.e. games developed for PS4/Xbox1) will be GPU-bound, and games ported to PC, using MANTLE, will be much more GPU-bound. As shown during APU13, a CPU as that in Kaveri will be able to feed the fastest graphics cards using MANTLE.

Resume:
=======

Old games: Kaveri =< i5-2500k

Modern games: Kaveri >= i5-2500k

MANTLE games: Kaveri ~ i7-4770k

Bonus:
=====

Interview with EIDOS. THIEF will come with improved multi-core support, MANTLE edition for "high performance graphics", and PhysX support. Enjoy!

http://www.dsogaming.com/interviews/eidos-montreal-talks-thief-tech-tessellation-fov-slider-multi-core-cpus-graphical-features/

 

[etayorius]

Nope, If I want play games at 1080p at medium settings testing CPUs at low settings and 720p is misleading. Moreover, as commented before, the tendency in games will be towards offloading the CPU generating GPU-bounded games.

Now you talk two different things. You can either:

A: Bench the CPU at the settings where the difference in CPU is the greatest

OR

B: Bench the CPU at settings where everyone actually plays at.

The problem with the second, of course, is any differences in CPU power is suppressed due to the GPU bottleneck in games at high settings. Running at lower settings shows the difference in power between two CPUs when said bottleneck is removed. Frankly, telling me CPU A and CPU B perform the same in game X with GPU Y is meaningless; what if I have a faster GPU, do the two CPU's still perform the same? Woops, don't have a clue. By testing at minimal settings though, you see which one is truly more powerful, even if the FPS numbers are meaningless.

Frankly, both ways should be done; min settings to show the order of CPU's in terms of power on said game, and at max settings to determine typical FPS with a given hardware setup.

Double review is a good idea, and some sites do that, finding that A10-6880k is ~40% behind i5-2500k at low settings and 720p/1024p, but only ~8% behind at high settings and 1080p. All this using a 7970 GHz Ed with Windows 7 and 'older' games. The differences are smaller with modern games under W8.1 or SteamOS.

Which is a somewhat typical result. GPU bottleneck at high settings hiding the CPU difference, which is at least 40% for that given game (possibly more, if you could get settings lower via GPU driver tweaks). Run at low settings where CPU power is FAR more important, then the differences in power (and power reserve) becomes clear.

Which infers that unless AMD increases performance by 40%, Kaveri won't be catching the 2500k, at least in that particular game.

So understand what review sites are benching: They typically bench how a given processor(s) perform at given settings in a given benchmark. They do NOT directly compare the relative processing power of said processors when doing these benchmarks. This leads to false impressions, like a 6880k being only 8% slower then a 2500k.

First, I wrote "games", plural. The above percentages are averages.

Second, this comparison was made under Windows 7 (which is optimized for Intel). AMD APUs are 5-15% faster under W8.1 or SteamOS that under W7, thanks to a new scheduler than understand the CMT architecture of AMD modules, closing the artificial gap with Intel chips.

Third, the comparison used 1600MHz RAM for both, when Richland stock memory speed is 2133MHz. Add another 5% by using stock memory timings for Richland.

Therefore, AMD increasing hardware performance by ~ 20% is enough to caught the i5 in so-called CPU benchmarks. The rest 10-20% comes from using stock memory and an operative system not optimized for Intel 😉.

Fourth, the result of the 6880k being only 8% slower than a 2500k at 1080p at high settings is real and measured. If you play those games at that resolution and settings, that is the performance than you obtain. It is not some imagined score.

Fifth, as explained again and again and again and again in this thread. Future games (i.e. games developed for PS4/Xbox1) will be GPU-bound, and games ported to PC, using MANTLE, will be much more GPU-bound. As shown during APU13, a CPU as that in Kaveri will be able to feed the fastest graphics cards using MANTLE.

Resume:
=======

Old games: Kaveri =< i5-2500k

Modern games: Kaveri >= i5-2500k

MANTLE games: Kaveri ~ i7-4770k

Bonus:
=====

Interview with EIDOS. THIEF will come with improved multi-core support, MANTLE edition for "high performance graphics", and PhysX support. Enjoy!

http://www.dsogaming.com/interviews/eidos-montreal-talks-thief-tech-tessellation-fov-slider-multi-core-cpus-graphical-features/

Man, if you`re right ill be running to the store as fast as i can to buy my Kaveri and a FM2+

 

[noob2222]

Mantle games: llano=trinity=i3 2100=kaveri=4350=2500k=g630=4770k in theory. Its purpose is to remove the cpu bottleneck so its a moot point to constantly bring it up. Not every game will be built with it so what measure is used then?

Mantle isn't just about "kaveri"

Do people base their purchases soley based on a few programs or overall performance.

Even so in those few programs any cpu will work, not just kaveri.