Wednesday, January 14th 2009

AMD Justifies Use of Large L3 Cache on Phenom II, Opteron

AMD's introduction of the Phenom II series processors served several purposes and goals for the company, mainly porting the processor technology to the newer 45nm SOI manufacturing node, to attempt to bring down manufacturing cost. This also meant that AMD could trade-off bringing down manufacturing cost with stepping up transistor counts on a die that is nearly the size of that of the 65nm Barcelona/Agena. The 45nm Shanghai/Deneb has a distinct feature over its predecessor: three times the amount of L3 cache. The larger cache significantly adds to the transistor count of the die: 758 million as against the 468 million on Barcelona/Agena. Replying to an inquiry of Hardware-Infos, AMD attempts to explain its motive behind incorporating the large L3 cache, while trading-off with savings of die-size and alleged latencies the L3 cache brings in.

AMD points out that expanding the L3 cache was important to the architecture in more ways than one. On the desktop/client PC front, the additional L3 cache was expected to provide a 5% performance increment over its predecessor. The reviews later backed AMD's assertion. Secondly, AMD likes to maintain an essentially common die design for both its client (Phenom II/Deneb) and enterprise or server (Opteron/Shanghai), to make sure manufacturing costs aren't wasted in setting up a separate manufacturing node. With the enterprise-grade Opteron processors, the 6 MB L3 cache has proven to benefit the processor in dealing with large server workloads. Finally, AMD claims that despite the larger cache, the overall die-area of the 45nm die remains lesser than that of the 65nm Stars die, so cost-cutting remains to an extant.

Source: Hardware-Infos
Add your own comment

28 Comments on AMD Justifies Use of Large L3 Cache on Phenom II, Opteron

#1
Melvis
eidairaman1 said:
all these parts are unlocked
I also have a old skt A system here at home, my GF uses it at the moment, its been a great old system, but i had to overclock the cpu for it to run stable =/

I didnt see much if any performance gain with DDR2 over DDR so thats why i kept with my old machine, and far as i can tell the faster the ram is the more you can overclock ya cpu? thats all i can tell that makes it worth wile...o and also its super cheap.

Id love to do up a skt A beast with a 3200+ and good ram and a 3850, it be sweet :D

I would of like to have seen AMD put alot of L2 cache on there new CPU's instead of L3 as my bro also says its just for show, but maybe not? it helped out C2D alot....
Posted on Reply
#2
eidairaman1
newtekie1 said:
The latency on good DDR2-800 RAM is the same as that on good DDR-400. DDR2-800@4-4-4-12 is the same latency time as DDR-400@2-2-2-6, those are actually pretty damn good latenncies for DDR. Even if you look at some of the best DDR, you can find DDR-500 with latencies of 3-3-2-8, and you can find DDR2-533 with 3-3-3-8, or even DDR2-1066 with 5-5-5-15, which actually has better latencies. People see the higher timings of DDR2 and 3, and don't take into consideration the higher clock speeds.
usually takes higher clock speed to get things done in DDR2/3 vs DDR, it takes 800 MTS DDR2/3 to get the same stuff done that DDR does at 400 MTS, Double Clock speed, thats just like how Rambus was at the time, imagine 800MTS DDR vs 800MTS DDR2/3, the DDR would clobber DDR2/3, Im pretty sure if they just shrunk the DDR1 Process to say 80nm it would impact DDR2/3 sales if DDR1 was still produced. DDR1 Maximum Speed was PC4800 (600MTS). To me they are just milking DDR technically now as there isnt much benefit like there was when PC1600 came about.
Posted on Reply
#3
newtekie1
Semi-Retired Folder
eidairaman1 said:
usually takes higher clock speed to get things done in DDR2/3 vs DDR, it takes 800 MTS DDR2/3 to get the same stuff done that DDR does at 400 MTS, Double Clock speed, thats just like how Rambus was at the time, imagine 800MTS DDR vs 800MTS DDR2/3, the DDR would clobber DDR2/3, Im pretty sure if they just shrunk the DDR1 Process to say 80nm it would impact DDR2/3 sales if DDR1 was still produced. DDR1 Maximum Speed was PC4800 (600MTS). To me they are just milking DDR technically now as there isnt much benefit like there was when PC1600 came about.
The higher clock speeds allow for more bandwidth, which was something that the old FSB system that Intel continued to use needed. The few Intel 775 boards that had DDR support, performed like crap with DDR compared to DDR2. It isn't about getting things done, the modern Intel processors needed more bandwidth, improving latency wasn't a high priority.
Posted on Reply