Guess what product AMD is revealing!

[Vesku]

If they were being forthcoming they would be talking about performance/watt, which is a metric that the industry and AMD itself recognizes as meaningful and is what they use in the press releases for their other products. Instead they are quoting performance/TDP, which is something rather useless, especially if the two TDP metrics aren't equal for AMD, something they didn't disclose. Plus one chip has turbo, the other does not. Depending on how aggressive turbo is on Beema and the testing conditions (starting temperature, cooling), that will really twist the results.

I'm sure that the even if Beema is more efficient than Kabini, the numbers will really fall short of reaching the numbers AMD is extracting with their performance/TDP mess. We'll know soon enough.

They do talk about performance per watt, just in very specific benchmarks with footnoted system specifications. The performance per watt slide claim is 2x in "3DMark11 - Performance" and "PCMark8 - Home" footnote (1) but Anandtech did not post the slide backup document so we do not know how comparable the Kabini and Beema systems were.

Piesquared posted that slide it can also be found in my Anandtech link.

AMD has been unusually specific in pre-releasing performance and performance/watt claims for this Cat family update. Imo, it is unlikely that they'd hand out slides that will completely embarrass them when the product finally arrives.

 

[Sweepr]

More than 2x the performance per watt of the current A4-1200 crap is not impossible with a refined chip + Turbo even on the same process. Trying to draw Silvermont vs Jaguar per/watt conclusions out of total desktop platform power was a funny joke though. Just a reminder, 2013 Atom Z3770 is already 3x faster than that slow 1GHz dual-core Jaguar CPU-wise (0.46 vs 1.48 @ CB 11.5 MT). They have half a year to convince an OEM to release a relevant Mullins-based tablet/convertible till Cherry Trail-T launches, blowing Bay Trail-T out of the water when it comes to graphics performance (16 Gen 8 EUs vs 4 Gen 7 EUs) and probably perf/watt too. Clock is ticking.

 

[piesquared]

And intel will have a half a year to 'convince' OEM's before Beema/Mullins replacement. Oh wow look at that, it works both ways!

 

[Vesku]

More than 2x the performance per watt of the current A4-1200 crap is not impossible with a refined chip + Turbo even on the same process. Trying to draw Silvermont vs Jaguar per/watt conclusions out of total desktop platform power was a funny joke though. Just a reminder, 2013 Atom Z3770 is already 3x faster than that slow 1GHz dual-core Jaguar CPU-wise (MT). They have half a year to convince an OEM to release a relevant Mullins-based tablet/convertible till Cherry Trail-T launches, blowing Bay Trail-T out of the water when it comes to graphics performance (16 Gen 8 EUs vs 4 Gen 7 EUs) and probably perf/watt too.

It has always been a race when it comes to computer products. That they are potentially announcing its, almost certainly 'soft', arrival next week and it has been said by AMD's head of Global Business that Beema has been shipping to customers for at least a few weeks if not several months (Q1 2014) means that they do have some breathing room before Intel's 'Trail' update arrives.

 

[Sweepr]

And intel will have a half a year to 'convince' OEM's before Beema/Mullins replacement. Oh wow look at that, it works both ways!

Just in time for Intel's next gen Goldmont CPU architecture + Gen 9 iGPU graphics, aka Broxton (14nm) then. Fun times ahead.
While Intel certainly needs to catch up in Android tablets/phones, AMD is playing catch up in the Windows mobile arena.

 

[Abwx]

http://www.anandtech.com/show/7514/amd-2014-mobile-apu-update-beema-and-mullins

AMD was unusually specific with their claims regarding Beema/Mullins:

Granted if they 'fudge' Beema TDP by a watt or two compared to how they rated Kabini that can magic up a chunk of their efficiency claims. Heck in the unlikely event that 15W TDP Beema draws the same actual power as 25W TDP A6-5200 it's still ~20% more efficient in perf/W at least in the two benchmarks AMD was willing to release. Impressive without a node shrink.

Perf/watt in PC Mark is 74.44 pts/Watt for Kabini and 154.13 pts/Watt for Beema, perfs of the lower power variants are 167.87 pts/Watt for Temash and 402 pts/watt for Mullins, if accurate thoses numbers point to a quite good process at GF where those chips are most likely fabbed.

Anandtech did not post the slide backup document so we do not know how comparable the Kabini and Beema systems were.

This was posted elsewhere, same plateforms for Kabini/Beema and for Temash/Mullins, i guess that some will be caught off guard if just changing the chip yield thoses improvements...

 

[piesquared]

Just in time for Intel's next gen Goldmont CPU architecture + Gen 9 iGPU graphics, aka Broxton (14nm) then. Fun times ahead.
While Intel certainly needs to catch up in Android tablets/phones, AMD is playing catch up in the Windows mobile arena.

Which will be just in time for AMD's products with stacked RAM (stacked RAM comes sooner?) and next gen GCN. Yep, still working both ways! Fun times indeed.

 

[mrmt]

They do talk about performance per watt, just in very specific benchmarks. So in some specific benchmarks it is 2x performance per watt.

Well, that would be cherry picking, no?

 

[Sweepr]

Which will be just in time for AMD's products with stacked RAM (stacked RAM comes sooner?) and next gen GCN. Yep, still working both ways! Fun times indeed.

Sadly, those are not coming in H2/2015, Broxton targeted launch. Ran out of arguments?
Also, where did you see the AMD part with stacked RAM? Cherry Trail-T and Broxton are both confirmed by now, what about this chip?
Intel couldn't care less about AMD in the tablet/smartphone/convertible space right now with giants like Qualcomm rulling the market and big vendors producing their own SoCs.

 

[Vesku]

Well, that would be cherry picking, no?

When has Intel and AMD not cherry picked? Cherry picking is basically the defacto in consumer electronics. Not sure how AMD could design it so that it is 2x in two popular benchmarks, one GPU and one a desktop use system suite, and not at least decently more efficient in general.

 

[mrmt]

When has Intel and AMD not cherry picked? Cherry picking is basically the defacto in consumer electronics. Not sure how AMD could design it so that it is 2x in two popular benchmarks, one GPU and one a desktop use system suite, and not at least decently more efficient in general.

Usually you can have a very good idea of the performance of Intel upcoming processors with the info they release in events and in previews. Sandy Bridge, Ivy Bridge, Haswell, Bay Trail, among others, are examples of that.

Can you say the same about AMD? Did Bulldozer live up the hype the company was making? What about Richland? Or Kaveri? And do you think a company with THIS track record should have any credibility, especially when using shady metrics like performance/TDP?

 

[rtsurfer]

Intel couldn't care less about AMD in the tablet/smartphone/convertible space right now with giants like Qualcomm rulling the market and big vendors producing their own SoCs.

Actually the Android phone space is a bit stagnated right right now.

I would love for Intel & AMD to come make some noise.

Samsung was rumored to come with a 64 bit SOC with the Galaxy S3 which didn't happen.
Then they tried to come up with an ARM BIG-LITTLE implementation in the S4 which was broken.
So much for their hyped Octa core.

Qualcomm has been re releasing the same Quad core SOC the they released 2 years ago as the Snapdragon S4 Pro.
Just manufacturing it on a different process & raising the Clock Speed.

The only one innovating in the SOC space is Apple. Which they keep limited to their crap products.
So Android Arena is pretty dry right now.

At this pace Intel & AMD might catch-up soon.

 

[Shivansps]

Well, AMD's TDP ratings with Kabini seem quite reasonable based on third party reviews. It would be somewhat surprising if they tossed that with Beema/Mullins, didn't they start using Intel's SDP for that kind of "trickery"? IIRC, that's how they were trying to pitch Temash for tablets, an SDP rating.

It's not as if OEM product design teams are fooled easily by such sleight of hand anyway. Cooling and psu/battery will have to work on the actual product not just on paper.

Edit: I did remember correctly. AMD is using SDP for Mullins which they also how they rated Temash but sticking with TDP, how Kabini was rated, for Beema.

"First, we have Beema replacing Kabini, and with the change we get the AMD Security Processor (ARM Cortex-A5) and a reduction in TDP on some parts, with 10W being the minimum. Mullins does the same for Temash, only AMD uses SDP (Scenario Design Power) rather than TDP (Thermal Design Power), and the new APUs are ~2W compared to 3-4W for Temash. Apparently the TDP for Temash is 8W and the TDP for Mullins is 4.5W, and that’s what AMD uses for their performance per watt calculations."

http://www.anandtech.com/show/7514/amd-2014-mobile-apu-update-beema-and-mullins

AMD was unusually specific with their claims regarding Beema/Mullins:

Granted if they 'fudge' Beema TDP by a watt or two compared to how they rated Kabini that can magic up a chunk of their efficiency claims. Heck in the unlikely event that 15W TDP Beema draws the same actual power as 25W TDP A6-5200 it's still ~20% more efficient in perf/W at least in the two benchmarks AMD was willing to release. Impressive without a node shrink.

Wait, AMD numbers are based on PC Mark 8, most likely accelerated, if the GPU its upgraded to GCN 2.0 can help a lot there whiout helping to the product value, the choice of 3Dmark11 instead of Fire Strike score is also fishy.

Im not sure if its fair to strap in a better GPGPU igp cores, run GPGPU accelerated benchmarks and call it twice perf/W.

 

[Vesku]

Wait, AMD numbers are based on PC Mark 8, most likely accelerated, if the GPU its upgraded to GCN 2.0 can help a lot there whiout helping to the product value, the choice of 3Dmark11 instead of Fire Strike score is also fishy.

Im not sure if its fair to strap in a better GPGPU igp cores, run GPGPU accelerated benchmarks and call it twice perf/W.

Ah PCMark 8 favors GPU Compute more than previous versions, yeah I'd say the main improvement is in the GPU section which is the section that seems to have visible changes.

 

[Shivansps]

GPU improvements will also decrease power, if i learned anything from A4-1200 reviews is that the GPU on Kabini is power hungry.
Thats probably where AMD is hidding the magic of twice perf/w, combined with GPGPU stuff, maybe even HSA support.

 

[monstercameron]

https://semiaccurate.com/forums/showpost.php?p=210953&postcount=287

The Dell specs listed 4 model numbers: E1-6010, E2-6110, A4-6210 and A6-6310. It was also rumored that AMD may release dual-core E1-6050 and E1-6100, together with quad-core E2-6200, A4-6300, A6-6400 and A8-6500 parts. As for tablet "Mullins" processors, AMD prepares three SoCs with a bit unusual model numbers: E1L-620T, A4L-640T and A10L-670T. The E1L-620T has 2 cores and 1 MB of L2 cache, while the A4L-640T and A10L-670T have 4 cores and 4 MB of L2 cache.

http://www.cpu-world.com/news_2014/...h_Beema_and_or_Mullins_core_on_April_29_.html

Model #: A8-6500, A8-6400, A6-6400, A4-6300, E2-6200, E1-6100, E1-6050, A10L-670T, A4L-640T, E1L-620T
Part #: AM6500ITJ44JB, AM6400ITJ44JBA, AM6400ITJ44JBA, AM6300ITJ44JBA, EM6200ITJ44JBA, EM6100IUJ23JBA, EM6050IUJ23JBA, AM670TIVJ44JBA, AM640TIVJ44JBA, EM620TIWJ23JBA

https://www.sata-io.org/sata-everyw...t_id=All&field_product_category_tid=39&=Apply

 

[AtenRa]

Usually you can have a very good idea of the performance of Intel upcoming processors with the info they release in events and in previews. Sandy Bridge, Ivy Bridge, Haswell, Bay Trail, among others, are examples of that.

You mean slides like the one bellow that shows 3x graphics performance over SB ??? :whiste:

And actual Gaming performance,

http://www.anandtech.com/show/5771/the-intel-ivy-bridge-core-i7-3770k-review/9

Although Ivy graphics was an improvement over SB, those 2-3x performance increase claims were only marketing or "cherry picked".

Can you say the same about AMD? Did Bulldozer live up the hype the company was making? What about Richland? Or Kaveri? And do you think a company with THIS track record should have any credibility, especially when using shady metrics like performance/TDP?

Pretty much YES,

AT Trinity review
http://www.anandtech.com/show/6332/amd-trinity-a10-5800k-a8-5600k-review-part-1

AT Kaveri Review
http://www.anandtech.com/show/7677/amd-kaveri-review-a8-7600-a10-7850k

On both occasions you can find the performance claims to be very close to real performance tested on all reviews around the net.

 

[OCGuy]

Hopefully something that can infuse some cash and give them a chance to be a real competitor again.

 

[Shivansps]

They changed the image today.

 

[erunion]

They do talk about performance per watt, just in very specific benchmarks with footnoted system specifications. The performance per watt slide claim is 2x in "3DMark11 - Performance" and "PCMark8 - Home" footnote (1) but Anandtech did not post the slide backup document so we do not know how comparable the Kabini and Beema systems were.

Piesquared posted that slide it can also be found in my Anandtech link.

AMD has been unusually specific in pre-releasing performance and performance/watt claims for this Cat family update. Imo, it is unlikely that they'd hand out slides that will completely embarrass them when the product finally arrives.

This has been thoroughly covered. AMDs efficiency claim is not based on measured power usage. They simply divided the benchmark score by nominal TDP. It says so in their disclaimer.

So no, we wont see 2x performance per watt and AMD technically never claimed we would. They used a different measurement(of dubious value)

 

[AtenRa]

They changed the image today.

thx

 

[Skurge]

Hopefully something that can infuse some cash and give them a chance to be a real competitor again.

I doubt they ever will be on the mid to high end desktop. Low end and mobile they may be a competitor.

 

[cbn]

Hopefully something that can infuse some cash and give them a chance to be a real competitor again.

If only AMD could make two lines of SOCs compatible with AM1.

1.) smaller cat cores (Jaguar, Puma ,etc) (Example: quad core - two wide OoO on a 25 watt SOC)

2.) large cat cores (Lion, Tiger, etc) (Example: dual core - three wide OoO cores on a 25 watt watt SOC)

That would keep me very interested (especially given the durability of current motherboard designs).

Furthermore, If the large cat core AM1 SOC could have four SATA, but still be able to work with existing AM1 boards with only two SATA ports.

 

[parvadomus]

We are almost there, cant wait to see what improvements this chip has over jaguar.

 

[blackened23]

If only AMD could make two lines of SOCs compatible with AM1.

1.) smaller cat cores (Jaguar, Puma ,etc) (Example: quad core - two wide OoO on a 25 watt SOC)

2.) large cat cores (Lion, Tiger, etc) (Example: dual core - three wide OoO cores on a 25 watt watt SOC)

That would keep me very interested (especially given the durability of current motherboard designs).

Furthermore, If the large cat core AM1 SOC could have four SATA, but still be able to work with existing AM1 boards with only two SATA ports.

Any particular reason you're talking about desktop parts? If these SKUs hit the desktop they will be nothing more than low performance low cost SOCs just as Kabini or desktop BT are. Kabini was a mobile part that somehow ended up as a super cheap and low performance desktop chip a long time later. Beema and Mullins are the same.....intended to be mobile products. They are not designed with desktop as a prime consideration. For desktop as a prime consideration you should look at Kaveri. Beema + Mullins - These products are designed for mobile first and desktop as a last consideration.

This is pointed out in AMD's roadmap, so seeing this talk of SATA is confusing. You shouldn't be worried about SATA or core count. You should be thinking about mobile metrics. The only metrics that are going to matter for Beema and Mullins are overall performance and the performance per watt. With kabini and temash in terms of mobile, they didn't fare so well there - the parts with appropriate battery life performed poorly, while the parts with better performance used too much power. Personally I think it will only be a slight improvement over Kabini and Temash which would not be a good thing. But we'll see. Maybe AMD pulled a rabbit out of the hat, but given their R+D situation I kinda doubt it. I'd have to think that it's going to only be a slight improvement just as Kaveri was compared to Richland. But, again, we'll see.