• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

AMD Carrizo APU Massive Leak – 40W TDP, Excavator Cores, 512 SP and Assorted Benchmarks

I agree. I keep hoping that when it comes to a rebuild that AMD will have something that can truly rival Intel.

The bank manager disagrees.

There's no money left in the enthusiast desktop market. That's why it's now completely half assed and server rebrands.

The money is in the mobile/tablet/phablet/flaptop :D market. And that's where AMD are fishing.

Intel are doing exactly the same. Shrinking dies so they can get into the market where the money is.

And that's just a sad fact I'm afraid. Expect to see more companies drop by the wayside next year.
 
So long as you need those bloody lanes there'll be a market.

An increasing amount of the die area of Intel consumer chips are the IGP and chipset functionality,and the focus is more and more on IGP performance and power consumption. They are spending billions of dollars on subsidising chips for tablets and improving graphics performance.

To put it in context Intel could easily put in 4 cores and 8 threads into the same die area as a Core i3 if they had no IGP and sell it for the same price.

The socket 2011 chips are the Intel server chips,but even that lasts as long as there is a market for large chips with multiple cores and no IGP. The AMD FX chips were their traditional server CPUs,but since they cannot compete on that market,I don't expect them to actually have a true successor to the FX CPUs for years.
 
Last edited:
Looks like AMD Carizzo will be 3.1 billion transistors in size and 244MM2:

http://www.eetimes.com/document.asp?_mc=RSS_EET_EDT&doc_id=1324643&page_number=2

AMD will disclose Carrizo, an integrated processor with its latest x86 core. The 28 nm chip measures 244.62 mm2 and packs more than 3.1 billion transistors. Its new Excavator core is 23% smaller and uses 40% less power than AMD’s previous x86 core.

Kaveri had 2.41 billion transistors and the same die area. I suspect the increased number of transistors is for the integrated FCH.

HDL used??
 
Excavator core is 23% smaller and uses 40% less power than AMD’s previous x86 core

Looks promising, if it uses the new memory compression it might still be a step up from 7850K etc, hopefully after Carrizo the APU's will get HBM.
 
40% less power on 28nm? When these are shrunk they should be an attractive little powerpack. I wonder what extra reduction you could get with manual design.
 
Looks like AMD Carizzo will be 3.1 billion transistors in size and 244MM2:

http://www.eetimes.com/document.asp?_mc=RSS_EET_EDT&doc_id=1324643&page_number=2



Kaveri had 2.41 billion transistors and the same die area. I suspect the increased number of transistors is for the integrated FCH.

HDL used??

AMD will disclose Carrizo, an integrated processor with its latest x86 core. The 28 nm chip measures 244.62 mm2 and packs more than 3.1 billion transistors. Its new Excavator core is 23% smaller and uses 40% less power than AMD’s previous x86 core.

Wow really? that would make an Excavator FX-8350 a 75 Watt CPU, which, yeah thats pretty good. bring it! :)
 
An increasing amount of the die area of Intel consumer chips are the IGP and chipset functionality,and the focus is more and more on IGP performance and power consumption. They are spending billions of dollars on subsidising chips for tablets and improving graphics performance.

To put it in context Intel could easily put in 4 cores and 8 threads into the same die area as a Core i3 if they had no IGP and sell it for the same price.

The socket 2011 chips are the Intel server chips,but even that lasts as long as there is a market for large chips with multiple cores and no IGP. The AMD FX chips were their traditional server CPUs,but since they cannot compete on that market,I don't expect them to actually have a true successor to the FX CPUs for years.

By all accounts you can get a Xeon that's basically a 4790 (non k) for the same price as a 4690k.

Loads of people on Linus' are lapping them up. Ed. 1230 v3
 
By all accounts you can get a Xeon that's basically a 4790 (non k) for the same price as a 4690k.

Loads of people on Linus' are lapping them up. Ed. 1230 v3

Yep and I had Xeon E3 CPUs for a while - in fact they were first noticed on US home server/VMWare threads and SFF PC forums,and IIRC I was one of the first to suggest on UK forums when one or two retailers started stocking them over here.

Even then half the die area of the chip is the IGP,so if Intel did ditch the IGP it would be probably Core i3 priced. You are paying a big cost for the IGP.
 
By all accounts you can get a Xeon that's basically a 4790 (non k) for the same price as a 4690k.

Loads of people on Linus' are lapping them up. Ed. 1230 v3

Yep and I had Xeon E3 CPUs for a while - in fact they were first noticed on US home server/VMWare threads and SFF PC forums,and IIRC I was one of the first to suggest on UK forums when one or two retailers started stocking them over here.

Even then half the die area of the chip is the IGP,so if Intel did ditch the IGP it would be probably Core i3 priced. You are paying a big cost for the IGP.

The Xeon's for are like a hidden gem, some don't seem to realize they will run in normal desktop mobo's and go for an i5 instead, losing the extra cache / threads at the same price, for those that don't bother with overclocking they are spot on. Used a Xeon 1230v2 for a while and it was a lil beast, ran cool and used hardly any power. The latest v3's are even better.

A non-overclockers dream :p
 
Sadly AMD will never truly hope to battle Intel until they compete on nanometers

I dont believe that is true. Architecture is an important thing and more so than the size of the process node imho. Look at the new samsung 540 pro ssds. Lightning fast and build on a larger 32(? not sure about the actual size) nm process yet it abselutely wipes the floor compared to most ssds, especially in the 4k read/write areas, all due to how the nand is designed(stacked).
 
Sadly AMD will never truly hope to battle Intel until they compete on nanometers

Depends on what you mean by "Battle Intel"

IMO AMD only need to get close to Intel on performance per Watt to be competitive.

If rumours are true;

AMD will disclose Carrizo, an integrated processor with its latest x86 core. The 28 nm chip measures 244.62 mm2 and packs more than 3.1 billion transistors. Its new Excavator core is 23% smaller and uses 40% less power than AMD’s previous x86 core.

They are very much on their way to achieving that with Excavator.
 
Carrizo and Carrizo-L announced:

http://www.anandtech.com/show/8742/amd-announces-carrizo-and-carrizol-next-gen-apus-for-h1-2015

http://techreport.com/news/27392/am...-2015-will-share-package-with-beema-successor

In an accompanying press release, AMD indicates that the first systems won't be available until "mid-year 2015."

So quite a while for actual systems though.

. A video by AMD’s VP/GM for Computing and Graphics, John Byrne, states that Carrizo and Carrizo-L are currently being tested internally ready for a 1H 2015 release, along with support for DirectX 12, OpenCL 2.0, Mantle and Freesync.

It supports DX12.

CN 1.2's lossless delta color compression algorithms help improve the performance in memory bandwidth limited scenarios, such as in APUs.

It seems that the next APUs will suppport memory bandwidth compression.

However people doubted me about what GPU changes AMD would make:

mupUpmD.png


One example of the efficiency improvement was provided by AMD’s Voltage Adaptive Operation. Rather than compensate for voltage variations which wastes energy, this technology takes the average operating voltage and detects when the voltage increases beyond a smaller margin. To compensate for this increase, the CPU speed is reduced until the voltage drops below the threshold and then the CPU speed is moved back up.

12L3mPx.png


The changes in speed are designed to be so minute that it does not affect overall performance, however it might only take an errant voltage delivery component to consistently make the voltage go above that threshold, causing erratic slowdown that might be statistically significant. It will be interesting to see how AMD implements the latest version of this feature.

So AMD is implementing similar features like in Nvidia Maxwell for its APUs meaning that its discrete GPUs will be getting similar tech too.
 
Last edited:
Back
Top Bottom