AnandtechAMD Carrizo ExcavatorReview

[Abwx]

Wait,what he was at AMD and they didn't even give him production representative laptops??

This is the day whatever little faith I had left in AMD marketing just died.

These were surely models dating from the Computex show, the "tests" clearly show that not only the idle power is abnormaly high, wich point to early bios and perhaps not even W10, and that the 15W chips are below 15W if we are to look at the deltas, even the MT tests do not reach this level..

 

[Phynaz]

Wait,what he was at AMD and they didn't even give him production representative laptops??

This is the day whatever little faith I had left in AMD marketing just died.

OH, it gets worse...

This unit was purchased especially for our testing (it turns out there’s a Best Buy around the corner from AMD HQ), and comes in low on the hardware all around.

AMD went and bought a craptop at Best Buy for testing.

 

[Shaun_Brannen]

I wonder who will be the last of the faithful. Even our resident AMD reseller is trashing AMD these days.

I don't think you can narrow it down to one... at this point, you've got your core "Jonestown" types.

Although IIRC, "our resident AMD reseller" sells Intel stuff too.

 

[ShintaiDK]

OH, it gets worse...



AMD went and bought a craptop at Best Buy for testing.

 

[Phynaz]

Exactly. These are the people that are being counted on to save AMD.

 

[Shaun_Brannen]

Exactly. These are the people that are being counted on to save AMD.

I don't know, it doesn't seem that weird to me. Desperate times and measures, you know?

 

[caswow]

this thread went down the drain 8 pages ago...:thumbsup:

 

[frozentundra123456]

I don't think you can narrow it down to one... at this point, you've got your core "Jonestown" types.

Although IIRC, "our resident AMD reseller" sells Intel stuff too.

Doesnt matter what he sells, his posts speak for themselves.

 

[Bryf50]

this thread went down the drain 8 pages ago...:thumbsup:

Yea, it almost makes me sad to see the forums in this state.

 

[The Stilt]

Did few "Linux Benchmark" runs and the results this test produces is complete rubbish. The results make no sense and are all over the place. I cannot replicate even remotely the same results between two runs

I used the script from here on Ubuntu 14.04.3 LTS x86_64. I removed the tests from the shell script which are not used on Anandtech Carrizo review. Aside from being completely erratic, some of the tests (NPB) only utilizes two threads even the rest of the test utilize all four like they are supposed to.

If someone want to look at the bench logs, they can be found here

 

[jhu]

Did few "Linux Benchmark" runs and the results this test produces is complete rubbish. The results make no sense and are all over the place. I cannot replicate even remotely the same results between two runs

I used the script from here on Ubuntu 14.04.3 LTS x86_64. I removed the tests from the shell script which are not used on Anandtech Carrizo review. Aside from being completely erratic, some of the tests (NPB) only utilizes two threads even the rest of the test utilize all four like they are supposed to.

If someone want to look at the bench logs, they can be found here

Could you do Blender benchmark under Linux using these files?

 

[Shaun_Brannen]

Doesnt matter what he sells, his posts speak for themselves.

Just to be clear, we're talking about AtenRa, right?

I've actually gained quite a lot of respect for him over the years, even if I don't always agree with him. He's willing to admit he's wrong when he is, but sees things from a different perspective from myself. And at any rate, he's way less of a dick than I've been, or some others have been.

He's the kind of AMD fan I could chat over a beer with.

Some other people around here, though... well... lock us in a room, and only one of us will be alive after 30 seconds...

 

[The Stilt]

Could you do Blender benchmark under Linux using these files?

15/25W/DDR-1600 = 20m 2.95s
35/42W/DDR-2133 = 15m 23.23s

Default settings.

OpenCL could not be tested since there is no display driver which could be installed. The drivers from AMD are blocked from installing since they are "bad of quality"

 

[el etro]

15/25W/DDR-1600 = 20m 2.95s
35/42W/DDR-2133 = 15m 23.23s

Default settings.

OpenCL could not be tested since there is no display driver which could be installed. The drivers from AMD are blocked from installing since they are "bad of quality"

Jhu, how do the results fits there?

 

[Rngwn]

Just looked around and found 3DMark benchmark on Carrizo. This one is the A10-8700P with dual-channel RAM and 35W TDP... as the benchmark says so.

http://www.3dmark.com/fs/5947217

Fire Strike: 1030

Remember that this is the R6 iGPU with 384SPs not the most powerful one that Carrizo has to offer. Looks like the 512SPs version would lies between Intel's GT2 and GT3e if there is no throtting. Not too shabby eh?

 

[DrMrLordX]

i am a mod :sneaky:

You don't say?!?!?

Exactly. These are the people that are being counted on to save AMD.

Nobody expects AMD marketing to save anything.

OpenCL could not be tested since there is no display driver which could be installed. The drivers from AMD are blocked from installing since they are "bad of quality"

Wait, is that under Linux? What version of Ubuntu are you running again? I think you may need to use one that defaults to kernel 4.2 or later so you get amdgpu.

Also, if you are Linux benching, you may want to try the Phoronix suite . . . it takes a lot of tinkering/dling to get all that stuff working, but it's quite comprehensive.

 

[The Stilt]

wait, is that under linux? What version of ubuntu are you running again? I think you may need to use one that defaults to kernel 4.2 or later so you get amdgpu.

14.04.3 lts.

 

[Markfw]

OK, this is posted as a mod. Some posters posted erroneous information. Then people corrected that. While the facts are all correct with the later posters, I urge (and should have infracted) those to be "nicer". In the future, my heavy mod hand may come out.Just so you know this activity i frowned on, but sometimes truth is a valid defense.

 

[Phynaz]

Nobody expects AMD marketing to save anything.

But we expect the corporate executives whom Anandtech was working with to save the company. It's these same executives that are responsible for not making sure AMD was put in the best light for this article.

 

[jhu]

15/25W/DDR-1600 = 20m 2.95s
35/42W/DDR-2133 = 15m 23.23s

Default settings.

OpenCL could not be tested since there is no display driver which could be installed. The drivers from AMD are blocked from installing since they are "bad of quality"

Thanks. A few more questions: which of the files did you use (1 car or 2 cars)? What were the sustained clockspeeds? Which version of Blender did you use?

 

[The Stilt]

Thanks. A few more questions: which of the files did you use (1 car or 2 cars)? What were the sustained clockspeeds? Which version of Blender did you use?

Blender 2.76b 64-bit, BMW27.blend (two cars facing opposite directions). I cannot tell the frequency very accurately, but it should have stayed at around 3.3GHz (35/42W) during the test.

EDIT: for the 15/25W result ~ 2500MHz.

 

[Dresdenboy]

But we expect the corporate executives whom Anandtech was working with to save the company. It's these same executives that are responsible for not making sure AMD was put in the best light for this article.

As you mentioned before, they still missed the 2.4X power efficiency goal. Could that have been their focus - to get closer to that figure than showing high performance?
How much power consumption increase can be expected from adding a second DIMM? Is it in the 1 or 2W range incl. the additional SOC NB/I/O power?

There is also this unlikely hypothesis: AMD marketing purposely let it look like its product got unfairly reviewed on a well known enthusiast site to activate evangelists among the enthusiasts spreading the word, that the products are better than shown. This get's the "word of mouth" tag, with all its implications.

 

[coercitiv]

As you mentioned before, they still missed the 2.4X power efficiency goal. Could that have been their focus - to get closer to that figure than showing high performance?
How much power consumption increase can be expected from adding a second DIMM? Is it in the 1 or 2W range incl. the additional SOC NB/I/O power?

How can you get closer to your power efficiency goal when using 5-10% less energy drops performance by 30% or more?

There is also this unlikely hypothesis: AMD marketing purposely let it look like its product got unfairly reviewed on a well known enthusiast site to activate evangelists among the enthusiasts spreading the word, that the products are better than shown.

Nope, this time I'll go with the more likely hypothesis: AMD marketing is professionally retarded.

 

[jhu]

There is also this unlikely hypothesis: AMD marketing purposely let it look like its product got unfairly reviewed on a well known enthusiast site to activate evangelists among the enthusiasts spreading the word, that the products are better than shown. This get's the "word of mouth" tag, with all its implications.

That worked out great the last time they tried that with Bulldozer. It worked so well that JFAMD was getting hate mail. On a side note people are way too emotionally invested in their CPUs.

 

[Dresdenboy]

How can you get closer to your power efficiency goal when using 5-10% less energy drops performance by 30% or more?

That depends on the equations used (not our simple comparison of max power at a task). According to the footnotes, AMD used the ETEC equation 1 given on page 10 here:
http://www.energystar.gov/ia/partners/product_specs/program_reqs/Computers_Program_Requirements.pdf

Then they divide some an SpecIntRate, Passmark and PCMark score by that number. Passmark is heavily influenced by IDIV latency, PCMark uses GPGPU, SpecIntRate probably runs well out of the L2 cache. But for their own tests, they usually used 2 DIMMs.

That worked out great the last time they tried that with Bulldozer. It worked so well that JFAMD was getting hate mail. On a side note people are way too emotionally invested in their CPUs.

At least he didn't try to let BD look bad. And I fully agree with your last statement. I think it is like sports for many.