The AMD Execution Thread [2007 - 2017]

Status
Not open for further replies.
The confusion is rampant...this test was done with an 4670k vs 7850k with a 280x GPU. As expected, the i5 came out well ahead. Is it because the 280x is unsupported, the driver isn't done, or both? Nobody seems to know and there is conflicting data all over the place.

~ Any R7 or R9 card will work in dual-graphics vs only 240/250
~ Only DDR3 will work in dual-graphics vs GDDR5 versions of supported cards will work just fine

http://www.hardwareheaven.com/revie...e-g1sniper-a88x-performance-with-r9-280x.html

"Performance with R9-280X
Of course an integrated GPU is only part of the APU story... we can add certain AMD GPUs to our system (E.g. R7-270) and combine the APU Hawaii cores and GPU for enhanced performance. Something we will cover in an upcoming article, with a new driver designed to enhnce Kaveri Dual GPU performance. Then there is the option to add a higher specification disctete GPU..."
 
The inference in the "45% faster" slide is the compatibility of dual graphics with 7850k and 290x...as both are specifically cited. This is confusing, disingenuous, and flat out WRONG according to the 2 sources I just found:

http://news.softpedia.com/news/AMD-...TrueAudio-and-Fluid-Motion-Video-416847.shtml

"The performance difference (percentage-wise) won't be very large on R9 cards, but the improvements gained by systems with, say, a Radeon R7 240 are substantial.

According to AMD's tests, the Radeon R7 240, alone, runs Bioshock: Infinite at 20.4 FPS (frames per second) and Tomb Raider at 19.4 FPS.

In a Kaveri system, though, Dual Graphics tech pools the GPU resources together and increases the FPS.

The rate goes to 40.2 in the former (95% faster) and 38.1 in the latter game (again, 95% faster). So if you build a Kaveri-based PC, it actually pays to fork up $80 / €80 more for a low-end video board."

http://wccftech.com/amd-kaveri-dual-graphics-works-ddr3-memory-based-radeon-r7-gpus/

Update - We have just received information form an AMD representative that the AMD Kaveri APU officially supports the Radeon R7 250 and Radeon R7 240 with dual graphics technology. There’s no memory type restriction and the APU would be able to work with both DDR3 and GDDR5 variants of these cards. Do note that no other Radeon R200 series GPUs are supported by Kaveri for dual graphics compatibility. Following is the reply:

~ So out of the models you listed only the R7 240 and R7 250 are officially supported.
~ 7750 is not and neither is R7 260X
~ So this has nothing to do with DDR3 or GDDR5
~ R7 240 GDDR5 as well as R7 250 GDDR5 should work just fine.
~ In summary: R7 250 and R7 240 are the models that are supported for Dual Graphics with Kaveri.

You're right, I mixed up the HD 7750 and R7 250 (still not used to AMD's new naming scheme).

I didn't mean to say that GDDR5 somehow precludes the use of Dual Graphics, just that it gives the dGPU a big advantage over the APU, which throws the balance off and makes Dual Graphics less efficient.

The R9-280X is a high-end card, too fast for Dual Graphics to make sense.
 
If this is true, any benchmark that isn't 7850k with 240/250 to determine MAX graphics power with this combo actually hurts AMD (i.e. quickly diminishing returns in benchmarks)

Anything over a $100 dGPU makes the 7850 look gimped in comparison to Intel. Anything above $275 in hardware (CPU/APU & dGPU) AMD needs to push their 8320 and 8350 pushing Mantle against Nvidia.

Below $275 in combined hardware possible to take share from both Intel CPU and Nvidia discrete
Above $275 in combined hardware really only pushing Mantle vs Nvdia discrete makes sense

All these AMD sponsored slides or demos against 4770k or using the 290x are just FUD and will come back to haunt them imo.

If I were AMD, I would push:

Kaveri vs i5 4670 (no discrete)...win AMD
Kaveri with R7 250 vs Intel with same card...win AMD
AMD 8320/8350 with 270x, 280x, 290, 290x Mantle optimized vs i5 4670k/i7 4770k with Nvidia...win AMD

Anything else is just bad for AMD or bad for the consumer who buys Kaveri expecting his 270x, 280x, 290, or 290x to run better than a similarly priced system running a sub $200 CPU.
 
If this is true, any benchmark that isn't 7850k with 240/250 to determine MAX graphics power with this combo actually hurts AMD (i.e. quickly diminishing returns in benchmarks)

Anything over a $100 dGPU makes the 7850 look gimped in comparison to Intel. Anything above $275 in hardware (CPU/APU & dGPU) AMD needs to push their 8320 and 8350 pushing Mantle against Nvidia.

Below $275 in combined hardware possible to take share from both Intel CPU and Nvidia discrete
Above $275 in combined hardware really only pushing Mantle vs Nvdia discrete makes sense

All these AMD sponsored slides or demos against 4770k or using the 290x are just FUD and will come back to haunt them imo.

If I were AMD, I would push:

Kaveri vs i5 4670 (no discrete)...win AMD
Kaveri with R7 250 vs Intel with same card...win AMD
AMD 8320/8350 with 270x, 280x, 290, 290x Mantle optimized vs i5 4670k/i7 4770k with Nvidia...win AMD

Anything else is just bad for AMD or bad for the consumer who buys Kaveri expecting his 270x, 280x, 290, or 290x to run better than a similarly priced system running a sub $200 CPU.

Honestly, I'm not sure this consumer exists. AMD isn't putting a lot of marketing effort behind Dual Graphics (understandably so) and most reviews have shown fairly poor results, even with low-end/mid-range cards. I doubt anyone would expect anything from Kaveri with a big dGPU.
 
I doubt anyone would expect anything from Kaveri with a big dGPU.

Probably not for dual graphics, no. But isn't there still the possibility that HSA/Mantle-optimized Kaveri get games/software where it could take on a role analogous to a dedicated PhysX card or some such? Thereby, potentially leveraging the processing power of the iGPU to enhance the performance of the (higher end) dGPU, regardless of dual graphics?
 
Honestly, I'm not sure this consumer exists. AMD isn't putting a lot of marketing effort behind Dual Graphics (understandably so) and most reviews have shown fairly poor results, even with low-end/mid-range cards.

The AMD slides seem to show a dramatic improvement for the 240/250 cards running in Dual Graphics:

"According to AMD's tests, the Radeon R7 240, alone, runs Bioshock: Infinite at 20.4 FPS (frames per second) and Tomb Raider at 19.4 FPS.

In a Kaveri system, though, Dual Graphics tech pools the GPU resources together and increases the FPS.

The rate goes to 40.2 in the former (95% faster) and 38.1 in the latter game (again, 95% faster)."

I doubt anyone would expect anything from Kaveri with a big dGPU.

This brings up my initial criticism messaging for the Kaveri launch...messy and confusing. AMD using a 290x coupled with a 7850k for the "45% faster" claim for Mantle...reviewers using everything from 270x-280x and claiming some improvement due to Dual Graphics...several reviewers unsure of memory limitations in Dual Graphics...the list goes on and on. The whole thing is one big confusing mess from a messaging standpoint.

If you want a clear message you write it, you deliver it, and you stay consistent. Perhaps the muddled message will help them sell Kaveri to the legions who buy 4670k? If this is their idea, that is dangerous because once these retail consumers find out it only scales well up to a $90 dGPU...there will be a backlash.

Hopefully this will get straightened out in short order...
 
Whereas Mantle was a great news, kaveri made me quite a bit gloomy about AMD position in the PC realm. Simply put AMD has no compliant solution on any segment on the market. Jaguar burns too much power for tablets. Steamroller are not competitive with Intel offering.
It is not only perfs that are responsible their positioning is IMO off.
For example kabini, it is either too big /expansive or lack GPU power. Lot conf I see use salvaged cheap version of the chip. Amd on one hand produces lots different GPU chips and not enough CPU ones. They should have gone for 2 different chips in my opinion, a 2 cores, 2 cus one backed by a 64 bit bus a more high and one with 4 cores 4 CUs backed by a 128bit bus. The former should be cheap to produce and have good enough yields being super tiny,the latter could have allow for gaming enabled laptop on the low/mid end, that conf would have been a poster child for for Mantle benefits.
Now Kaveri is a tougher one as it is in no way competitive as be it for desktop or laptop a late Pentium + a cheap discrete GPU is away better investment. To make things worse they made it pretty big...
 
Last edited by a moderator:
http://www.tomshardware.com/reviews/a10-7850k-a8-7600-kaveri,3725-8.html

I think the conclusions are starting to gel...Kaveri will be relegated to the low end desktop market and will have more legs in laptops. Good product but hardly revolutionary...
Speaking of conclusion the one in Tom hardware review is pretty harsh
Personally, I’d go for a $75 Athlon X4 740 and a Radeon HD 7750 for close to the same amount of money…at least until HSA-optimized software gives us a reason to favor Kaveri-based APUs.
That is pretty awful if you ask me, even though he speaks of trinity based chips there should be replaced by salvaged Kaveri.
pretty much kaveri is intersting at less than 80$ with its GPU disable which might take more than half of the die. Nonsensical, AMD please wake up you are wasting money you don't have :???:
 
It just goes to show that the CPU in the APU is what wins people over. AMD's CPU is really awful. Their fancy GCN GPU is so bandwidth starved that it has limited value. Plus these chips are likely to look a lot worse when the TDP drops. I'm sure it will be just as ugly as with Llano and Trinity. Plus, the cheap notebooks they go into will probably continue the tradition of inadequate cooling that causes inevitable throttling if you fully utilize that TDP.

Also, it sure seems like Intel has GPU tech almost as figured out as NV and AMD at this point. If they spend more transistors on GPU and use the EDRAM, AMD probably can't win anymore.
 
AMD Sued for Overestimating APU Success to Investors

Some investors in AMD, who purchased company stock between October 27, 2011 and October 18, 2012, filed a class action lawsuit against AMD, through law firm Robbins Geller Rudman & Dowd. The class action alleges that AMD and its officers and directors committed violations of the US Securities Exchange Act of 1934. AMD allegedly misled its investors over how popular its first generation "Llano" desktop APU would get, claiming that there were much greater prospects for the APU than it actually ended up selling.

The law firm detailing its suit writes:
The complaint alleges that during the Class Period, defendants made false and misleading statements about the Company's business and prospects. Specifically, the complaint alleges defendants made false statements and/or concealed adverse facts regarding AMD's 32 nanometer Llano (the "Llano") Accelerated Processing Unit ("APU"), which is a type of microprocessor that combines AMD's central processing unit and its graphics processing unit onto a single piece of silicon, including repeatedly highlighting the "strong" and "significant" interest in, demand for, and unit shipments of, the Llano APUs, and falsely and misleadingly representing that AMD's desktop business was in a "strong position" and that it would "continue to rebound" in 2012. As a result of defendants' false statements, AMD stock traded at artificially inflated prices throughout the Class Period.
 
Eh, AMD has historically been somewhat slimy when it comes to what it tells investors and consumers, partly because it's had more periods where it was embarrassing itself. At least those watching AMD's delay-fest with Llano had reason to suspect there was something less rosy about it.
I'm pretty sure AMD was at least a little sketchy with Llano, but I don't recall a class action over Barcelona or R600, and that was way worse.
 

Not sure it is entirely the fault of the GPU...

Forcing the machine to use Intel integrated graphics also seemed to help the problem until the system went down after about one day of normal use.

There's a chance that increased heat generation inside the MacBook Pro due to having an additional dedicated graphics card may be accelerating a problem that exists in the design. IE - replace with any other heat generating part and it's possible you'd still end up with the same series of problems.

Regards,
SB
 
I am upping my 4Q 2013 numbers. I am adding an extra 1 - 1.5 million console units vs my earlier predictions which were already up 1 million units. I am also going to add another $100 million for dGPU sales...which may be low...but there is something going on there although it is VERY difficult to measure.

I was at $1.65 billion...now $1.85 billion
I was at .11 cents...now .14 cents

Looking for $5.00 - $6.00+ this week after earnings.

Or I could be living in a box under the freeway come Tuesday...lol
 
This is a red herring...has no legs...and is a sign of the large hedge funds being overly concerned about their short position. It has far more to do with overestimating the health of the desktop market than overly rosy promises about APUs...specifically Llano.

Your big bold red "headline" is misleading, wrong, and FUD.

wow

Actually, I do prefer to support these legal actions against the company... Perhaps it would help them from ethical and moral standpoint to improve all their strategies, actions, and work as a whole... Because the company offers underwhelming products after underwhelming products...

http://vr-zone.com/articles/amd-faces-class-action-investor-lawsuit-llano-sales/69944.html

During AMD second quarter earnings call, the company announced that weak demand for Llano had caused the company to report lower than expected revenue for the quarter. The next trading day, AMD’s stock dropped 25 percent on this news. Months later in October 2012 AMD announced that it was taking a $100 million writedown on unsold Llano inventory, which could not be moved to channel because of a weak demand for the product in China and Europe. Subsequently, AMD’s stock dropped 17 percent the following trading day.

That hurts a lot the owners of shares
 
Changing socket from Llano to Trinity screws up the chances of selling older stock for cheap.
No one would ever want a Llano now, but it could have been a decent option for low-ends even today.
 
http://vr-zone.com/articles/amd-faces-class-action-investor-lawsuit-llano-sales/69944.html

During AMD second quarter earnings call, the company announced that weak demand for Llano had caused the company to report lower than expected revenue for the quarter. The next trading day, AMD’s stock dropped 25 percent on this news. Months later in October 2012 AMD announced that it was taking a $100 million writedown on unsold Llano inventory, which could not be moved to channel because of a weak demand for the product in China and Europe. Subsequently, AMD’s stock dropped 17 percent the following trading day.

That hurts a lot the owners of shares

From the same article you linked:

"This case might not be as cut and dry as the plaintiffs and their class action suit chasing law firm hope for. During the last few years the PC market has been in an overall downward trend, with demand also softening for components of AMD’s rival Intel. While a jury (the plaintiffs demand a jury trial) may be convinced that the language used by AMD executives was deceptive, AMD’s counsel has a strong case to argue that their company was simply victim to an industry-wide trend."

It simply won't amount to a hill of beans. Everybody got surprised by the 2012-2013 PC downturn...including Intel.
 
Status
Not open for further replies.
Back
Top