Talk:Athlon 64 X2
From Wikipedia, the free encyclopedia
Contents |
[edit] Cool'n'Quiet
"One issue that is not immediately obvious until you start using a dual core or processor system is that Cool'n'Quiet is somewhat incompatible with performance. Cool'n'Quiet works by checking how active the processor is while running and adjusting the speed and voltage up or down appropriately. If the processor is idle it slows it down, and if it is busy it speeds it up. With X2 cores this is actually detrimental to performance as both cores are not treated independently by Cool'n'Quiet but rather as a single unit. In the typical case where one core is busy and the other is idle you end up with a bizarre ping pong effect as the cores are continuously sped up and slowed down, resulting in an overall performance loss rather than the expected gain. To reap optimum performance one has to turn Cool'n'Quiet off and forgo the benefits in power saving and heat reduction."
I have CoolnQuiet enabled and constantly monitor my processor speed and voltage both in windows and linux. I have yet to experience this "ping-pong"-effect.
- I removed it, it was unsourced and if I remeber correctly Cooln' q has no effect on performace. Martin 15:44, 19 October 2005 (UTC)
-
- In theory it doesn't. CnQ reduces processor speed when it's not being used to it's maximum so it has no actually effect on performance. However if what the above says is true in that the dual cores are treated as a single unit then it's possible I guess that they could both be slowed down when one is close to maximum use resulting in performance issues. However I doubt it and as you said, it's unsourced Nil Einne 01:19, 2 June 2006 (UTC)
[edit] Picture
Hmm. The picture says that it is of a 3800+ Athlon X2 E6. However, the text says that all Manchester cores are E4. If all 3800+'s are E4, then how is this picture possible? --Roguelazer 03:35, 29 December 2005 (UTC)
-
- Either Toledo with partially disabled L2-Cache, typically too much defects to be sold as 2x1MB Toledo so it's sold as 2x 512KB Manchester or AMD is moving their complete Athlon64 line to E6 stepping. --Denniss 04:10, 29 December 2005 (UTC)
[edit] Development time
- Due to the nature of their application, most 3D games cannot be effectively multithreaded without disproportionate development time. There are exceptions, the most famous being the Quake III engine on which r_smp = 1 may be set, enabling multithreading support. The benefit is slight, but present.
This is a bit confusing. I guess what's it's trying to say is that 3D games can't be effectively multithreaded without disproportionate development time so most games are not multithreaded. The Quake III engine is multithreaded however it's not really done that well/effectively multithread (or perhaps it's not possible to do it well) so it doesn't provide that much of a benefit. I.E. Some time was spent making the engine multithreaded but not as much as is necessary for it to work very well (but more time then most other games.
At the moment, it sounds like it's saying the Quake III engine was effectively multithreaded without a disproportionate development time Nil Einne 01:15, 2 June 2006 (UTC)
While this is probably true, could some provide a reference? Also, perhaps this needs to be clarified. Is it really always going to require a lot of development time or could it be partially just that game programmers have never had a need to make multi-threaded game so they haven't and so learning how to do it is going to take a long time but once they start to get the hang of it, it'll be a lot simpler? Nil Einne 01:02, 2 June 2006 (UTC)
- I agree, it should be clarified a bit, and sourced - it seems to me that it's a bit of both, that it's been unnecessary and difficult, so multithreading hasn't been implemented in most games, but this should be clarified a bit, as the way it is worded now implies that it is nigh-impossible to multithread a game, which seems to be exaggerated. It should also be stated that games may benefit from additional cores, even though the game is not multithreaded, from things like background processes and other operations that are not necessarily part of the game. Fiskars007 00:12, 27 July 2006 (UTC)
[edit] Dual core responsiveness
Many users I've seen have commented on the improved responsiveness of a computer in normal, everday multitasking use. This isn't really mentioned/clear since it only suggests intensive multitasking scenarios and only performance (not responsiveness) which perhaps makes it seem like your decompressing an archive, while compressing a video and playing a game all at normal priority (or maybe having 50 FireFox tabs/windows along with Photoshop, Acrobat, Illustrator open)... Of course, this may be related to Windows scheduling code but it needs to be mentioned. Nil Einne 01:07, 2 June 2006 (UTC)
[edit] Mention of Energy Efficient X2's?
the energy efficient X2s.. eg 3800+ @ 35W..?
I was under the impression that the 35W X2's had different model numbers. This should be added in another paragraph - the model numbers and technologies should be researched and sourced. Also, isn't the 35W feature part of Cool'N'Quiet (which also seems to be absent from the article since the edit mentioned at the top of this talk page)? Fiskars007 00:12, 27 July 2006 (UTC)
[edit] Multi-core pros/cons
Removing this paragraph, as it's not specific to the Athlon 64 X2, it's about multi-core processors in general: Stevage 04:29, 27 October 2006 (UTC)
- Due to the nature of their application, most 3D games cannot be effectively multithreaded without disproportionate development time. There are exceptions, the most famous being the Quake III engine on which r_smp = 1 may be set, enabling multithreading support. The benefit is slight, but present. It's left to easily parallelizable tasks (image manipulation, media encoding, etc.) where different parts of the workload can be worked on independently to properly extract the power of a multi-core processor.
[edit] Product reviews
At newegg, many of the product reviews state that the Intel version is superior. I checked, but they are also $100 more. Is the Intel equivalent of the same price actually superior? --Can Not 15:49, 12 February 2007 (UTC)
- Talk pages are not for general discussion on the article's topic. You might consider visiting a forum such as Anandtech. — Aluvus t/c 01:12, 13 February 2007 (UTC)
- I was just thinking that if the acusations were true, it should be mentioned. It could simply be "Intel Fanboy bantur", but I wasn't sure.--Can Not 23:05, 15 February 2007 (UTC)
-
-
- Prices are super volatile at the moment. AMD knows that Intel's got the better chip now, so they're cutting their costs to keep up in sales. What I don't know is if they've dropped them enough to real price/perfomance parity with the Core 2 Duos. Also, what does the DC in the recently announced 64 X2 DC 6000+ chip mean? 66.36.131.228 10:30, 20 February 2007 (UTC)
-
