amd got lazy so they just disabled the core of a phenom cpu like a 9950 w/c is 125w tdp. and since it only utilizes 2 cores, its way cooler, so they manage to decrease the tdp by a level with a slight increase in clock speed.
1 - 19 of 19 Posts
Quote:
Correct, the kuma is a agena core (PhI) with two cores disabled. And it should be about 20% faster clock for clock.
The reason the TDP is higher, is that while the other electronics there are disabled, the power still flows through them, so it has a higher TDP than the brisbane. The addition of the L3 cache adds to the power requirement as well.
Originally Posted by Dopamin3 You have to realize its a different architecture. Even though it is clocked slower and has the same amount of cache, it is faster than the older ones. It's the newer phenom architecture for dual cores called kuma vs the older brisbane. I believe it is roughly 10-20% faster clock for clock than the brizzys. |
The reason the TDP is higher, is that while the other electronics there are disabled, the power still flows through them, so it has a higher TDP than the brisbane. The addition of the L3 cache adds to the power requirement as well.
6
Thanks all! I got some awesome replies. Responses are below.
Quote:
wat?
increasing clock speed increases TDP. I already knew it was a Phenom with 2 cores disabled. I said that in mah post! ;-P thx tho!
Quote:
This was inexplicative. See above response to gerikoh's quote. thx tho!
Quote:
Dude. FTW. Thank you.
Quote:
That's what I figured.
![]()
It makes sense given the added shared L3 and the faster HT.
Quote:
Beautifully written. Thx so much! I did not know about the clock-for-clock speed ratio between the two CPU's, but I expected exactly what you said about power still flowing thru the cores. What a shame that they couldn't just break their circuits... not that simple tho. The power distro is complex, and intended to deliver 125 W to 4 cores, etc.
Quote:
Originally Posted by gerikoh amd got lazy so they just disabled the core of a phenom cpu like a 9950 w/c is 125w tdp. and since it only utilizes 2 cores, it way cooler, so they manage to decrease the tdp by a level with a slight increase in clock speed. |
increasing clock speed increases TDP. I already knew it was a Phenom with 2 cores disabled. I said that in mah post! ;-P thx tho!
Quote:
Originally Posted by sLowEnd The Kuma is actually a Phenom with 2 disabled cores That's why. |
Quote:
Originally Posted by Dopamin3 You have to realize its a different architecture. Even though it is clocked slower and has the same amount of cache, it is faster than the older ones. It's the newer phenom architecture for dual cores called kuma vs the older brisbane. I believe it is roughly 10-20% faster clock for clock than the brizzys. |
Quote:
Originally Posted by nemesis82 kuma > any other non phenom x2 of amd |
Quote:
Originally Posted by logan Correct, the kuma is a agena core (PhI) with two cores disabled. And it should be about 20% faster clock for clock. The reason the TDP is higher, is that while the other electronics there are disabled, the power still flows through them, so it has a higher TDP than the brisbane. The addition of the L3 cache adds to the power requirement as well. |
2
Quote:
what i mean is, it's supposed to be 125w. but since they disabled 2 cores, it was reduced to 95w. the added mhz from the original 2.6ghz x4 9950 is just a bonus.
![]()
Originally Posted by Slink wat? increasing clock speed increases |
2
Quote:
Oh, righto, righto. It's kind-of-a gimmick (extra MHz) cuz it's B.E. anyway.
Originally Posted by gerikoh what i mean is, it's supposed to be 125w. but since they disabled 2 cores, it was reduced to 95w. the added mhz from the original 2.6ghz x4 9950 is just a bonus.
|
Quote:
man its hotter coz its a 95w tdp on stock processor compared to brisbane's 65w tdp..in average when you OC both the kuma can reach to 100w tdp and with the brisbane at 70w or more only..the higher TDP processor of course will generate more heat..
peformance wise the KUMA is afvantageos because of the extra L3 cache and the SSE4 intructions sets which brisbane lacks..
furthermore higher HT frequency gives more bandwidth for data transfers a kudos to kuma for this although your latency will be slow in this case Brisbane wins over Kuma..
Originally Posted by Slink Why does the Kuma run hotter than the X2_5400_Brisbane_BE? The Kuma core (the dual-core phenom) boasts a 3600 MHz hypertransport over the 5400's "pale-by-comparison" 2000 MHz, and the presence of a 3 MB L3 cache (compared to the 5400 having no L3 cache) but all other specs are rather identical, including the 65 nm manufacture tech, the stock frequency, the identical L1 & L2 caches (2x128KB L1 & 2x512KB L2). Also note this post: http://www.overclock.net/amd-cpus/47...al-core-2.html What gives?! Is the extra thermal output related to the "disabled Phenom cores" of the Kuma? Thanks, -Slink |
peformance wise the KUMA is afvantageos because of the extra L3 cache and the SSE4 intructions sets which brisbane lacks..
furthermore higher HT frequency gives more bandwidth for data transfers a kudos to kuma for this although your latency will be slow in this case Brisbane wins over Kuma..
Quote:
righto, that's all well-said, but we've covered most of that. Of what "latency" do you speak? That of RAM?
Originally Posted by kairi_zeroblade man its hotter coz its a 95w tdp on stock processor compared to brisbane's 65w tdp..in average when you OC both the kuma can reach to 100w tdp and with the brisbane at 70w or more only..the higher TDP processor of course will generate more heat.. peformance wise the KUMA is afvantageos because of the extra L3 cache and the SSE4 intructions sets which brisbane lacks.. furthermore higher HT frequency gives more bandwidth for data transfers a kudos to kuma for this although your latency will be slow in this case Brisbane wins over Kuma.. |
i made a review of what exactly your inquiring about just digging up the docs in here..just wait a sec..
![]()
Quote:
latency is how fast your memory refreshes..but i think kuma doesn't benefit having low latency since it has a L3 cache for buffer
Originally Posted by Slink righto, that's all well-said, but we've covered most of that. Of what "latency" do you speak? That of RAM? |
Quote:
its a 50/50 chance to get a low latency..low timings does not affect latency that much..since the memory controller is built in the chip not in the chipset..so in other words its how will the processor do the refreshing..
Originally Posted by Slink REPPY REP REP. Yeah, I will have decently low latency anyway. Just bought some wickedsick ram. If the tight timings don't yield great benefit with the Kuma, I'll loosen them and ramp up the "speed" setting (hopefully, after a BIOS update...). ;-) |
Quote:
Oh, I SEE!
Here is my latency report from latency.exe from cpuz:
Code:
How does that look? ;-P heh.
Originally Posted by kairi_zeroblade its a 50/50 chance to get a low latency..low timings does not affect latency that much..since the memory controller is built in the chip not in the chipset..so in other words its how will the processor do the refreshing.. |
Here is my latency report from latency.exe from cpuz:
Code:
Code:
Cache latency computation, ver 1.0
[URL=http://www.cpuid.com]www.cpuid.com[/URL]
Computing ...
stride 4 8 16 32 64 128 256 512
size (Kb)
1 3 3 3 3 3 3 3 3
2 3 3 3 3 3 3 3 3
4 3 3 3 3 4 3 3 3
8 3 3 3 3 3 3 3 3
16 3 3 3 3 3 3 3 3
32 3 3 3 3 3 3 3 3
64 3 3 3 3 3 3 3 3
128 4 6 8 16 17 15 12 12
256 4 6 8 16 17 12 13 13
512 4 6 8 16 17 12 13 16
1024 4 7 12 26 49 108 110 114
2048 4 7 14 26 49 108 109 113
4096 5 7 14 26 51 109 111 122
8192 4 7 14 26 49 110 113 120
16384 4 7 14 28 49 110 112 116
32768 4 7 14 26 50 110 113 119
2 cache levels detected
Level 1 size = 64Kb latency = 3 cycles
Level 2 size = 512Kb latency = 13 cycles
now lets find a way to unlock it
![]()
large cache for the kuma will put that latency issue behind and will perform better overall
![]()
and no you cannot unlock kuma lol
1 - 19 of 19 Posts
