I've an Samsung Engineering Sample that unfortunately is under NDAQuote:
Originally Posted by situman
Printable View
I've an Samsung Engineering Sample that unfortunately is under NDAQuote:
Originally Posted by situman
What does Samsung have to do with AMD's IMC latency?Quote:
Originally Posted by nn_step
it has more to do with the memory latency than has to do with AMD's IMC ;)Quote:
Originally Posted by situman
you need the memory to perform well if you want AMD's IMC to give any descent performance
1. learn what is MCQuote:
Originally Posted by nn_step
2. learn why MC is integerated on K8
3. learn how MC is integrated on K8
4. learn what DRAM/synchronous DRAM is
5. learn what SRAM is
6. learn on which latencies K8 performance depends the most
7. learn what cache is
8. learn what is exclusive and what is inclusive cache
9. learn what kind of softwares have temporal/spatial locality for data/instructions
10. learn how to measure SRAM latency
Doesn't that sort of apply to Intel as well? Lower DDR2 latencies generally benefits both platforms no? Also, your argument was that AMD's IMC latency is decreasing, so how did memory get into this? We already know that DDR2 latencies are dropping as evidenced by some of the forum members here. But, can you provide backup saying that AMD's IMC latency is decreasing? What Samsung is doing has nothing to do with AMD's IMC latency.Quote:
Originally Posted by nn_step
LOL :rofl:Quote:
Originally Posted by LOE
how many times do I have to tell you it isn't the basement :mad: it is the CAD Lab :p:Quote:
Originally Posted by LOE
Actually nnstep, conroes L2 cache latency is 14ns, but dothans was 7, and banias's was 6, yonahs is 12.
now that is music to my ears. :D
Can I have an example, please:rolleyes:Quote:
Originally Posted by LOE
right ;)Quote:
this is a discussion forum, not a smart ass forum right?
I was not arguing, and AMD R&D team in the basement can't give remote knowledge to nn. He has to learn something before he can discuss about it.Quote:
and don't you dare argue with nn - don't you know the AMD R&D team is working at his basement? ;)
source here
http://www.hkepc.com/bbs/attachments...h49MNBORXI.jpgQuote:
Sources indicate that AMD plans to release Athlon 64 X2 3600+ earlier from Q4 to Q3, probably few weeks after the price-cut on July 24. Targeting on Intel Pentium D 945 (3.4GHz/2MB x2 L2/800MHz FSB/No VT), Pentium D 915 (2.8GHz/2MB x2 L2/800MHz FSB/No VT), and the upcoming Core 2 Duo E4200 (1.6GHz/2MB/800MHz) and Pentium D 925(3.0GHz/2MB x2 L2/800MHz FSB), X2 3600+ is expected to priced in $130~149.
As we reported before, AMD Athlon 64 X2 3600+ is a cheaper version of 3800+, both clocked at 2GHz, but halved the L2 Cache to 256KB x2. According to the internal test from manufacturers, 3600+ is a little bit weaker than Pentium D 945, similar to Core 2 E4200, but faster than Pentium D 925/915. It will be a big obstacle to clear up Pentium D if 3600+ got a right price.
Besides, to avoid self competing with Conroe, Intel has to sell the stored Pentium D as soon as possible. Therefore, Conroe is set at a less competitive price. The appear of AMD Athlon 64 X2 3600+ would definitely lower the speed of clearing old Pentium D, helping AMD to fight the battle until Fab 36 and 3rd party manufacturing in Singapore be ready.
Figure below is a AMD Athlon 64 X2 3600 Socket AM2 version with OPN Code AD03600IAA4CU, where ADO stands for 65W TDP, 3600 for the model name, I for Socket AM2, AA for Variable V_Core and Variable T_Case, and 4CU for 512KB L2 (256KB x2) Stepping F4.
Are those Conroes in the background? :DQuote:
Originally Posted by StyM
… and it isn’t anywhere on these roadmaps:
http://www.dailytech.com/article.aspx?newsid=3170
BTW
gOJDO has again wrote complete OT, and his post should be deleted! So Perkam please do what admin/mod does!
Sempron64 X2
No socket F :rolleyes:Quote:
Originally Posted by gOJDO
they are LGA775 ;)
you can't count the pins from that angle :rolleyes:Quote:
Originally Posted by gOJDO
It is not socket F 100% and it is LGA775 100%
The shape, the size and the label in the middle.
http://www.stickied.org/files/c4d646...e37391a04c.jpg
Quote:
Originally Posted by gOJDO
And how this has ANYTHING to do with the thread title???
it is on the picture:Quote:
Originally Posted by informal
http://www.hkepc.com/bbs/attachments...h49MNBORXI.jpg
So what?Is Core2duo produced by AMD :confused: ???I guess not.
Why would you offtopic the subject with close up of the Core2,instead of talking about the 3600+ X2???
I was wondering why there are other chips in the background of the picture. Can you explain it please?
The subject of the topic is not about the other chips.It's about the 3600+ X2.
BTW ,it's HKEPC photo,so they can take a picture of an elephant coupled with a X2 3600+,it's their choice since it's their website.
Let's stay on topic here,and stay away from intel for once in a while.OK?
We don't need Core2 threadcrapping in AMD topics all over.It's a great chip,going to be fast etc etc.Let's talk about X2 3600+ possible effect on market,if it's going to be widely available etc.
Damn, why people have low expectations on me. I was curuois, WTF Conroe is doing in the backgorund. I thought there was a reason why it is in the background and that is all.
About 3600+, it will be cheap and it will probably be a good overclocker. It will take less diespace than singlecore with 1Mb L2, but it will offer better performance. It will be the best bang for the money from AMD and maybe we'll see higher clocked variants latter.