AMD Carrizo APU

AMD’s Carrizo APU A10-8890K CPU-Z Leaked – Six Excavator Cores with 95W TDP on the 20nm NodeNode

AMD-20nm-Carrizo-APU-A10-8890K.png
 
Última edição:
Se a plataforma dos FXs durou muitos anos a dos APUs muda a cada ano!

FM2 > FM2+ > FM3? Que loucura.

Mas é interessante ver já APUs 6 cores.

EDIT: Dizem que é fake!
 
Demonstraram hoje um sistema carrizo ao público.

http://anandtech.com/show/8855/amd-demonstrates-working-carrizo-laptop-prototype

Só ficámos a saber que tem hardware decoding de HEVC/H265, mais nada...

One of the features of Carrizo is full support for H.265 decoding, and as an example of why this is needed they had an Intel system running next to the Carrizo system attempting to playback a 4K H.265 video. While the AMD system was easily able to handle the task without dropping any frames, the Intel system was decoding at what appeared to be single digit frame rates. The 4K content was essentially unwatchable on Intel.
 
Mais ao menos aquilo que estava à espera.

AMDKFD Linux Driver Begins Preparing For VI Carrizo APUs

Volcanic Islands is the codename for the AMD GCN 1.2 architecture that's currently in use by just the Radeon R9 285 "Tonga" graphics card. GCN 1.2 provides a more efficient ISA, video scaler improvements, a new multimedia decode/encode engine, improved performance, and other optimizations.
http://www.phoronix.com/scan.php?page=news_item&px=MTg4MzU

a info é retirada do próprio patch para o novo open source driver para linux

This patch-set starts to prepare amdkfd so it could support VI APU.
http://lists.freedesktop.org/archives/dri-devel/2015-January/075093.html

EDIT: quer dizer este amdkfd é apenas uma parte do novo driver, é para implementar o HSA, o open source driver é o amdgpu


Em relação ao H265 decode, review da Anandtech à 285 e ao update ao UVD (Unified Video Decoder)

GCN 1.2 - Image & Video Processing
http://www.anandtech.com/show/8460/amd-radeon-r9-285-review/4
 
Última edição:
A 285 não tinha H265 decode nessa review, ou já lá estava no hardware (UVD 6.0) e estão a implementá-la agora no software?

Também gostava de saber se o VCE 3.1 tem suporte para H265 e se melhorou na qualidade além da velocidade (12 streams de 1080p ao mesmo tempo é muita fruta...)
 
O mais que te posso dizer é que o UVD e o VCE são implementados via IP da Tensilica (entretanto adquirida pela Cadence), agora qual os modelos específicos usados em cada versão, isso não sei.

Xtensa processors configured to accelerate video stream decoding are an ingredient in every UVD-powered AMD ATI graphics chip. UVD-powered AMD Radeon graphics chips include the HD7000, HD6000, HD5000, HD4000, HD3000, HD2000, x1900, x1600, and the x1300 series members, as well as chips containing the AMD Avivo technology.
http://ip.cadence.com/about/customer-profiles/graphics

Mas pode dar-se o caso de acontecer aquilo que disseste, a capacidade do hardware estar lá mas não estar implementada via driver, algo que o autor da review também deixa em aberto por a norma H265 ainda não ter as especificações finais definidas à data do artigo.

A AMD fez algo parecido com o TrueAudio por exemplo.
 
Dado que se trata apenas de optimização de arquitectura e nada realmente radical era mais do que esperado. Aliás, isto tem sido optimizações atrás de optimizações desde o Bobcat (bobcat->piledriver->Jaguar->Seamroller->Puma->Excavator) sem nunca ter havido uma mudança radical no funcionamento dos CPU's. O jaguar e o Puma ainda trouxeram melhorias simpáticas de IPC mas agora já se está a chegar ao ponto que espremer mais não vai trazer milagres.
 
Última edição:
Dado que se trata apenas de optimização de arquitectura e nada realmente radical era mais do que esperado. Aliás, isto tem sido optimizações atrás de optimizações desde o Bobcat (bobcat->piledriver->Jaguar->Seamroller->Puma->Excavator) sem nunca ter havido uma mudança radical no funcionamento dos CPU's. O jaguar e o Puma ainda trouxeram melhorias simpáticas de IPC mas agora já se está a chegar ao ponto que espremer mais não vai trazer milagres.

Estás a trocar tudo. Bobcat, Jaguar e Puma são baseados na arquitectura low power da AMD.
Bulldozer -> Trinity -> Richland -> Kaveri -> Carrizo são baseados nas melhorias da arquitectura Bulldozer.
A AMD tem ap+ostado constantemente na eficiência porque não consegue grandes ganhos de performance com cada geração. Basta olhar para o A10-7800K e para o A8-7600 para se perceber isso. Um consome muito mais energia que o outro e as diferenças em benchmarks são muito pequenas.

Por isso é que a AMD nem lança estes APUs para desktops ou CPUs baseados nesta arquitectura.
 
Epá, tenho de deixar de fazer posts em estado zombie :-D

Tens razão, eu misturei todas as linhas. Neste caso será então:

Bulldozer -> Piledriver -> Steamroller -> Excavator e vai-se lá saber o porquê ontem aluceinei tudo junto com a linha bobcat->Jaguar->Puma->Puma+ lol.

Seja como for o meu ponto mantêm-se pois tirando a passagem para 28nm do steamroller e algumas optimizações de arquitectura (mais no bulldozer -> piledriver) não houve nada revolucionário nos cores.

É mesmo esperar que a arquitectura Zen traga novamente a AMD para o topo (mas sinceramente não estou convencido que tal aconteça)
 
Estás a trocar tudo. Bobcat, Jaguar e Puma são baseados na arquitectura low power da AMD.
Bulldozer -> Trinity -> Richland -> Kaveri -> Carrizo são baseados nas melhorias da arquitectura Bulldozer.
A AMD tem ap+ostado constantemente na eficiência porque não consegue grandes ganhos de performance com cada geração. Basta olhar para o A10-7800K e para o A8-7600 para se perceber isso. Um consome muito mais energia que o outro e as diferenças em benchmarks são muito pequenas.

Por isso é que a AMD nem lança estes APUs para desktops ou CPUs baseados nesta arquitectura.

É uma realidade, e é pena ser assim pois quem ganha com isto é a Intel. É preciso lançar uma nova arquitectura, mas só lá para 2016 (arquitectura Zen).
 
http://wccftech.com/amd-carrizo-bas...ealed-fx8800p-leads-pack-stoney-ridge-family/

Hassan Mujtaba
AMD has detailed their upcoming APU codenamed Carrizo which will be featured inside mobility platforms. The Carrizo APU family comes in two series, one powered by the powerful x86 Excavator cores while the other APU codenamed Carrizo-L is powered by their x86 PUMA+ core architecture. Yesterday, along with the ISSCC details, we also covered a leak from 3DMark which was related to the flagship chip of the lineup, known as FX-8800P. Now, it can be confirmed that the FX-8800P is indeed the flagship APU of AMD’s Carrizo lineup along with the other models that will be available at launch in mid-2015.



AMD Carrizo APU “Stoney Ridge” Family Unveiled – A-Series, FX-Series, R-Series APUs Detailed
The upcoming APU lineup is interesting since it’s only mobile/embedded based with the new A-Series, FX-Series and R-Series chips. So far, we know of 10 chips that will be making their way to mobility/embedded platforms in mid-2015 however there can be even more models planned for launch later. One thing is confirmed that the chips will be based on a mix of Carrizo and Carrizo-L specifications and since we have no first hand information on that available at the moment, we can’t tell what exact core configuration can be found on each model. The FX-8800P will be the flagship of the lineup which is currently the only APU which we know some specifications about. So the lineup includes the following models:

  • AMD FX-8800P
  • AMD PRO FX-8800B
  • AMD A10-8700P
  • AMD A8-8600P
  • AMD A6-8500P
  • AMD PRO A10-8700B
  • AMD PRO A8-8600B
  • AMD PRO A6-8500B
  • AMD RX-418GD
  • AMD RX-216GD
As detailed yesterday, the last flagship APU that was part of AMD’s Kaveri family was codenamed FX-7600P and the FX-8800P seems to be its successor from the Carrizo APU family. The 3DMark benchmark is vague on specs and details but it does list the FX-8800P as a 12 Compute Cores enabled part which confirms that the Carrizo APU on mobility will feature a maximum of 4 x86 Excavator cores and 8 GCN Compute units forming into the 12 Compute Cores which AMD has been branding since the launch of Kaveri (because the can be used as GPGPU). The testing was performed on the AMD Gardenia platform which is an internal codename for the test board that is used to run Carrizo samples. Rest of the specifications include the Radeon R7 graphics branding for the 512 SPs iGPU and a preliminary clock speed of 1.7 GHz base and 2.1 GHz boost. The chip scores around 2645 points in 3DMark 11 performance mode which can be compared to the 2150 (average performance) score of the Kaveri based FX-7600P APU.

So what else is in the lineup? AMD has two FX-Series APUs which are the top models. These include the FX-8800P and PRO FX-8800B for professional use. On the A-Series front, we have 6 new chips, the AMD A10-8700P, AMD A8-8600P, AMD A6-8500P, PRO A10-8700B, PRO A8-8600B and PRO A6-8500B. Surprisingly, we are also looking at some R-Series APUs based on the Excavator core architecture which might be part of the “Merlin Falcon” family. The two APUs codenamed RX-418GD and RX-216GD might feature the x86 Excavator core architecture as anticipated. For full details on Carrizo, you can visit the briefly detailed article here.

AMD R-Series “Merlin Falcon” SOC

The Merlin Falcon SOC will be based off AMD’s next generation Excavator core architecture and will feature up to four x86 Excavator cores with 4 MB shared L2 cache. This series will target the high performance SOC segment and can be thought of as the embedded version of AMD’s upcoming “Carrizo” APU. Merlin Falcon SOC will feature 2nd generation HSA support (HSA 1.0) along with Dual channel 64-bit DDR3 channel (ECC). The higher performance SOC will feature the PCH on the same die that will allow PCI-e Gen 3 and Gen 2 along with SATA, USB 3.0/ 2.0, SPI, LPC support. The next generation graphics chip will allow for DirectX 11.1 graphics and a UVD 6 decoder. Some of the key features of Merlin Falcon SOC are highlighted below:

  • AMD Radeon HD 10000 Graphics architecture.
  • 4 Excavator CPU cores with 4mb of shared L2 cache.
  • 2nd generation HSA.
  • ARM security processor. Similar to the one featured in AMD’s latest Beema/Mullins APUs.
  • Dual channel DDR3 memory controller.
  • UVD (Universal Video Decode) 6 and VCE (Video Compression Engine) 3.

Read more: http://wccftech.com/amd-carrizo-bas...leads-pack-stoney-ridge-family/#ixzz3Sdp0yU14

AMD R-Series “Merlin Falcon” SOC
AMD-Embedded-Solutions-2012-2016-Roadmap-635x356.jpg

The Merlin Falcon SOC will be based off AMD’s next generation Excavator core architecture and will feature up to four x86 Excavator cores with 4 MB shared L2 cache. This series will target the high performance SOC segment and can be thought of as the embedded version of AMD’s upcoming “Carrizo” APU. Merlin Falcon SOC will feature 2nd generation HSA support (HSA 1.0) along with Dual channel 64-bit DDR3 channel (ECC). The higher performance SOC will feature the PCH on the same die that will allow PCI-e Gen 3 and Gen 2 along with SATA, USB 3.0/ 2.0, SPI, LPC support. The next generation graphics chip will allow for DirectX 11.1 graphics and a UVD 6 decoder. Some of the key features of Merlin Falcon SOC are highlighted below:

  • AMD Radeon HD 10000 Graphics architecture.
  • 4 Excavator CPU cores with 4mb of shared L2 cache.
  • 2nd generation HSA.
  • ARM security processor. Similar to the one featured in AMD’s latest Beema/Mullins APUs.
  • Dual channel DDR3 memory controller.
  • UVD (Universal Video Decode) 6 and VCE (Video Compression Engine) 3.

Read more: http://wccftech.com/amd-carrizo-bas...leads-pack-stoney-ridge-family/#ixzz3Sdp6Xogb
 
Back
Topo