ajax
Banido
Se calhar nunca ouviste falar de cache trashing.
AMD "Barcelona": clocks
- Autor do tópico blastarr
- Data Início
Se calhar nunca ouviste falar de cache trashing.
blastarr
Power Member
"Cache Thrashing".
E ocorre em qualquer processador, especialmente naqueles que possuem mecanismos de prefetching menos inteligentes, e com menor espaço para armazenar os dados.
A estrutura e quantidade das cache L1 e L2 dos futuros cores K8L é basicamente a mesma que existía nos Athlon XP "Barton"... (L1) (64+64) + (L2) 512KB.
Última edição:
Zar0n
Power Member
N n era de homem, pk nesse caso a cache L2 ia ser mais lenta e mt mais difícil de controlar.
Na verdade no K8L a L2 vai ser mais rápida já k é privada k a L2 da Intel, mas tds os dados vão sendo puxados para a L2, e em aplicações multicore vai ser mais rápido e escalar melhor já k os dados n vão passar pelo FSB.
Uma solução nativa é sempre superior, pk é k a Intel fez uma com o conroe? e aposto k a proxima arquitectura deve ser nativa com 4 e mais tarde 8 colados....
E quanto mais cores adicionas pior vai ser para a cache partilhada...
Errado a cache é non-inclusive.
Como ha uma cache L2 dedicada no K8L a cache partilhada n vai sofrer tanto como a L2 dos C2D/C2Q logo menos thrashing, mais um ponto para o K8L.
Uma das vantagens do conroe é um prefetching mais evoluido k anula em conjunto com a maior cache o controlador de memoria incluído.
Mas o K8L vai ter um prefetching mais agressivo em conjunto e Out of order loads e maior buffer.
As velocidades n são mt evevadas, mas sendo o 1º 65nm da nova arquitectura e para servidores podem estar a ser cautelosos com os yields.
Pela margem de OC pode-se logo ver a capacidade do 65nm nesta arquitectura, n se podem fazer comparações com os actuais 65nm K8, pois na altura o processo estava menos evoluído, quando chegarem os desktop a maturidade do processo deve permitir vel. mais elevadas.
Última edição:
Koncaman
Utilizador Saloio
isto esta a começar a ser uma conversa demasiado recorrente.
eu acho que não devemos olhar para a cache L3 pela sua velocidade, mas sim pela sua funcionalidade. Na minha opinião, a cache L3, não se destina propriamente a armazenar dados com alta probabilidade de virem a ser usados.
para mim, a cache L3 funciona (ou esta la para funcionar) mais como um buffered bus "intercore", do que como uma memoria cache propriamente dita, e acho que podemos dar como dado adquirido, que sempre que um core precisar de dados de qualquer outro, vai ter esses dados sempre com o mesmo tempo de espera, coisa que não acontece com o kentsfield, que contrabalança as coisas com 2 caches L2 partilhadas por cada par de cores.
no fundo resta saber se o (1--2) (3--4) do kentsfield funciona melhor que o (1---2---3---4) do Barcelona...
lembrem-se é que nos core2 a L2 tem que fazer igualmente de cache... e daí que a intel esteja interessada em aumentar a sua dimensão
neste momento, a teoria parece-me inconclusiva, podemos descobrir vantagens pelos 2 lados.
mas isto é falando a nivel de cache, não é apenas na cache que existem novidades... e eu acho que não é a cache que proporciona aos core 2 duo, os resultados que os temos visto obter...
ajax
Banido
Pois não mas a questão não é essa. É que para além da cache, controlador de memória, etc. os core 2 duo e os "K8L" são semelhantes em matéria de "novidades".
ajax
Banido
Parece que houve pessoas que não perceberam o comentário que fiz quando disse que a concorrência (Intel) era fraca no que diz respeito aos Quads. Deixo apenas um comentário do próprio Hector Ruiz.
HECTOR RUIZ: "In spite of all the hype and hoopla, there is no really such thing that Intel has leapfrogged AMD. It's quite the contrary. As a matter of fact, despite all the perceptions of Intel closing the gap, half the time they do a little bit better and the other half we do. And all of that will end with the introduction of Barcelona because it's such a significant jump in performance and quality."
HECTOR RUIZ: "In spite of all the hype and hoopla, there is no really such thing that Intel has leapfrogged AMD. It's quite the contrary. As a matter of fact, despite all the perceptions of Intel closing the gap, half the time they do a little bit better and the other half we do. And all of that will end with the introduction of Barcelona because it's such a significant jump in performance and quality."
Esperemos bem que o Hector Ruiz não esteja a "cuspir para o ar" pois eu vou comprar PC novo lá para o fim do ano, e quanto mais renhida
tiver a questão entre AMD/Intel, melhor...
Pois só assim é que evoluímos a um ritmo natural, e não ao sabor de uma qualquer política empresarial unilateral.
Ou será que já é isso que está a acontecer agora, só que ao ivés de ser unilateral é concertada?! What is the MATRIX?? Dasse...que estes comprimidos dão um coice....
mascarilha
I folded Myself
cumps
Última edição:
DJ_PAPA
Power Member
Mais um artigo explicando as mudanças da arquitectura ponto por ponto:
Barcelona Architecture: AMD on the Counterattack
http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2939&p=1
Recomendado.
+ News:
Basicamente o que eles indicaram:
- quad-core Barcelona 42% mais rapido em floating point que o topo de gama actualmente da intel. O quad-core Xeon X5355.
- E em integer performance tem uma vantagem sobre o quad-core Xeon de 2 digitos, isto é, >10%
Mesmo usando o Barcelona mais rapido (2.3Ghz) e bater o Xeon5355 a 2.66Ghz por aquela margem parece promissor. E ainda bem para ver se o mercado mexe para a frente.
Barcelona Architecture: AMD on the Counterattack
http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2939&p=1
Recomendado.
+ News:
http://www.eetimes.com/news/latest/showArticle.jhtml?articleID=197700269
Basicamente o que eles indicaram:
- quad-core Barcelona 42% mais rapido em floating point que o topo de gama actualmente da intel. O quad-core Xeon X5355.
- E em integer performance tem uma vantagem sobre o quad-core Xeon de 2 digitos, isto é, >10%
Mesmo usando o Barcelona mais rapido (2.3Ghz) e bater o Xeon5355 a 2.66Ghz por aquela margem parece promissor. E ainda bem para ver se o mercado mexe para a frente.
Última edição:
Zarolho
Power Member
Barcelona Architecture: AMD on the Counterattack
When Intel launched its first Core 2 based microprocessors, the performance improvement was beyond revolutionary. It was the biggest single performance improvement we had seen from a new microprocessor in several years at that point. A large part of Core 2's success was its architecture, but you cannot ignore that it couldn't have come at a better time for Intel.
All Conroe, Merom and Woodcrest had to do was outperform Intel's aging and misguided NetBurst based Pentium 4 processors. Doing so proved quite easy for AMD, which had been doing just that pretty much since the 2000 launch of the processors. With no competition from within Intel, AMD wasn't really doing that much better. While the K8 was a strong architecture, it was getting old. Without any serious performance enhancing architectural updates since its introduction back in 2003, AMD left Intel with a stationary target to aim for. With each successive iteration of the Pentium M architecture, Intel came closer and closer to developing its own Athlon 64 killer, eventually culminating in the release of such a product - the Core 2 Duo.
It wasn't some mystical force of microprocessor design prowess that allowed Intel to pull ahead last year; it was a good architecture and excellent timing. Ironically enough, it was the same two elements that orchestrated much of the success of AMD's K7 and K8 architectures; they were both good designs released at times when the competition was at its worst.
In terms of actual product releases, the first incarnation of AMD's new architecture will be found in the next-generation Opteron due out at the middle of this year. AMD will initially launch at speeds ranging from 2.1GHz to 2.3GHz, but by the end of this year you can expect higher clock speeds.On the desktop, AMD's Agena core will be a Barcelona equivalent shipping at between 2.7 - 2.9GHz. Kuma will be a dual-core variant of Agena shipping in the 2.0 - 2.9GHz range.
Barcelona will be a success for AMD; the long awaited architectural update to K8 should yield significant performance improvements, especially in current areas of weakness for the K8 (e.g. video encoding). Our review of the Athlon 64 X2 6000+ showed that with aggressive pricing, AMD could come close to offering something competitive to Intel. At current prices, we suspect that Barcelona would be enough to close the gap between AMD and Intel. Chances are that we won't see bargain basement prices on AMD's new cores, but we'd expect that AMD would have to maintain a competitive market.
The real catch here is what happens after Barcelona; as we mentioned before, Intel's current successes were born out of steady but regular evolution of a good starting architecture. With yearly updates to the Pentium M, Intel achieved a snowball effect that proved difficult to stop. It would seem that a similar approach by AMD would be necessary to avoid sticky situations like the one it finds itself in today.
http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2939
When Intel launched its first Core 2 based microprocessors, the performance improvement was beyond revolutionary. It was the biggest single performance improvement we had seen from a new microprocessor in several years at that point. A large part of Core 2's success was its architecture, but you cannot ignore that it couldn't have come at a better time for Intel.
All Conroe, Merom and Woodcrest had to do was outperform Intel's aging and misguided NetBurst based Pentium 4 processors. Doing so proved quite easy for AMD, which had been doing just that pretty much since the 2000 launch of the processors. With no competition from within Intel, AMD wasn't really doing that much better. While the K8 was a strong architecture, it was getting old. Without any serious performance enhancing architectural updates since its introduction back in 2003, AMD left Intel with a stationary target to aim for. With each successive iteration of the Pentium M architecture, Intel came closer and closer to developing its own Athlon 64 killer, eventually culminating in the release of such a product - the Core 2 Duo.
It wasn't some mystical force of microprocessor design prowess that allowed Intel to pull ahead last year; it was a good architecture and excellent timing. Ironically enough, it was the same two elements that orchestrated much of the success of AMD's K7 and K8 architectures; they were both good designs released at times when the competition was at its worst.
In terms of actual product releases, the first incarnation of AMD's new architecture will be found in the next-generation Opteron due out at the middle of this year. AMD will initially launch at speeds ranging from 2.1GHz to 2.3GHz, but by the end of this year you can expect higher clock speeds.On the desktop, AMD's Agena core will be a Barcelona equivalent shipping at between 2.7 - 2.9GHz. Kuma will be a dual-core variant of Agena shipping in the 2.0 - 2.9GHz range.
Barcelona will be a success for AMD; the long awaited architectural update to K8 should yield significant performance improvements, especially in current areas of weakness for the K8 (e.g. video encoding). Our review of the Athlon 64 X2 6000+ showed that with aggressive pricing, AMD could come close to offering something competitive to Intel. At current prices, we suspect that Barcelona would be enough to close the gap between AMD and Intel. Chances are that we won't see bargain basement prices on AMD's new cores, but we'd expect that AMD would have to maintain a competitive market.
The real catch here is what happens after Barcelona; as we mentioned before, Intel's current successes were born out of steady but regular evolution of a good starting architecture. With yearly updates to the Pentium M, Intel achieved a snowball effect that proved difficult to stop. It would seem that a similar approach by AMD would be necessary to avoid sticky situations like the one it finds itself in today.
http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2939
Última edição pelo moderador:
gafa
Power Member
Leitura muito interessante, explica bastante bem os caminhos da Intel e da AMD em termos de evolução e o funcionamento/diferenças do quad core "nativo" Barcelona por oposto aos c2d.
Só os quero é ver em acção, quero é benches para tirar ilações "concretas". Estou -farto- lol das especulações tanto com a r600 como com o Barcelona.
Cumps,
G
mascarilha
I folded Myself
cumps
