Friday, July 11th 2008
Pre-release Tests Conducted on AMD Denreb
Pre-release Tests Conducted on AMD Deneb
AMD Deneb is the code-name for the 45nm quad-core CPU which AMD plans to release soon. Chinese website ITOCP got their hands on two engineering samples. They used these samples at various clock-speeds set by altering the FSB multiplier and Vcore voltage. These chips were then subjected to rounds of Super Pi 1M benchmark. The results look rather luke-warm compared to what we saw of the Intel Bloomfield chips recently. The Deneb CPUs were supported by an AMD RD790 motherboard and 2 GB of DDR2 800 MHz unganged memory, running at timings of 5-5-5-18. The Phenom X4 Deneb 45nm will feature 6 MB of L3 cache apart from the usual 512 KB L2 caches dedicated to the cores.
Source:
ITOCP
AMD Deneb is the code-name for the 45nm quad-core CPU which AMD plans to release soon. Chinese website ITOCP got their hands on two engineering samples. They used these samples at various clock-speeds set by altering the FSB multiplier and Vcore voltage. These chips were then subjected to rounds of Super Pi 1M benchmark. The results look rather luke-warm compared to what we saw of the Intel Bloomfield chips recently. The Deneb CPUs were supported by an AMD RD790 motherboard and 2 GB of DDR2 800 MHz unganged memory, running at timings of 5-5-5-18. The Phenom X4 Deneb 45nm will feature 6 MB of L3 cache apart from the usual 512 KB L2 caches dedicated to the cores.
164 Comments on Pre-release Tests Conducted on AMD Denreb
Moving on - As far as the socket longevity argument between AMD and Intel, they are both just as guilty. The practices of the past have nothing to do with what will happen in the future. And for an example of AMD screwing people on a socket change, look no further than 939. Recap/summary = both have screwed customers with socket changes AND that does nothing to prove what will happen in the future.
Going back a few pages - My 45nm Quad does 4Ghz on 1.4v 24/7. Just giving an idea of the voltage differences between the 2 companies on the same process. I know mine doesn't represent any kind of real price/performance value.
Now, before anybody labels me a fanboy, know that I do still have my 6400+ running in my second rig, and that I still build AMD machines for clients when it represents the better value (for the non-OC'ers). It's just that 90% of the time, the Phenom doesn't represent the better value. The exceptions being when there's a special deal going on a Phenom, or the client uses heavily multi-threaded software on a regular basis, but doesn't want to spend the money on a higher end Phenom/Intel quad (Then the lower end Phenoms will do), or they already have a good AM2 board that supports Phenom (Hell, if my AM2 board supported Phenom, I would pick one up as a quick upgrade). All clients interested in OCing automatically get Intel systems tho.
When and if AMD ever takes the performance lead again, I will jump ship right back to them, until then, it's Intel for me.
i have to say even 1st gen AM2 boards "support" 45nm phenom but manufacturers wont update the BIOS on them they want you to spend YOUR money on THEIR new boards thing about that is its still faster than going to the ram 1.8ghz vs 1066mhz hmmm which do you think is faster
Intel never said that C2Ds, or Core 2 Quads will work in any LGA775 board. They always said that only certain boards support it so look at the manufacturer's documentation. Thats a better way to do business.
So socket support in recent years, was not so good with AMD.
And some of it's common sense. Don't buy a shitty board *cough* ECS (as someone mentioned) *cough*
I have nothing against AMD. I would buy their processors if they were priced more appropriately.
it intels chipset strategy
I also may juice the crap out of it to make the NB clock like I want to lol.
K
Like I was bitching earlier, it's the main perf. limitation of these chips.
I don't own one of the Intel quads (though like I said earilier I do like the QX6700 model) but they seem to pump less volts into their processors than AMD is doing at the time of course I imagine this is to play on the energy savings kick that everyone is on about these days.
I've been doing the energy savings things since I can remember so I do lots of testing with Phenoms and find that not a single one actually needs the voltage that they come with from the factory in order to run at stock speed...and sometimes even higher.
I'm running my Phenom 9850BE as we speak at 2.6Ghz (9950BE level) on 1.152v worth of juice 100% stable (I run roughly 10-12hrs worth of stress to test things) and it'll actually boot up at far far less than that voltage.
Doing this drops my power consumption by a HUGE degree down to the same levels that the Intel quads enjoy. My total system power consumption as I type this out, listen to music, have a few monitoring programs open, and all the other various programs associated with normal use is a whopping 151W.
I just ran POV Beta 3.7 to keep things fair and the max consumption rose to 213W at this setting. That's far lower than what the 9950BE is using at it's stock level and of course I could take it back to normal 9850BE speed but I was going after a certain point.
I'm sure that the Intel chips could go lower (not sure how much as they seem to be running low already :laugh:) to lower their consumption as well but I just wanted to show that the Phenoms aren't really as power hungry as they appear. AMD just for some reason pumps huge amounts of voltage through them anyway. Also note that my normal usage was at the same wattage levels that the 9950BE will get when Cool N Quiet is enabled (same with the Intels) and those were done at 1 hr idle time.
Crazy what a processor can do if you take the time to find out...
K
Credit goes to Legit Reviews for the power consumption figures of the other processors. For my readings I used the same parameters and tests that they did to achieve load numbers. I didn't bother taking an idle test however because you see my normal usage is what they get at idle already.
K
K
It´s not "DENREB"
it´s "DENEB"