ronch wrote:JBI, I'm in inclined to think it has something to do with SATA or som heavy I/O. (snip) For some reason, in retrospect it seems even moderately heavy I/O can stall a whole machine down a tiny bit regardless of the number of coars. Kinda makes the idea of having more cores moot, doesn't it?
Captain Ned wrote:Boy, I remember the bad old days when even burning a CD meant you started the burn then gently crept out of the room so as not to offend the spirits of burning.
Forge wrote:ronch wrote:JBI, I'm in inclined to think it has something to do with SATA or som heavy I/O. (snip) For some reason, in retrospect it seems even moderately heavy I/O can stall a whole machine down a tiny bit regardless of the number of coars. Kinda makes the idea of having more cores moot, doesn't it?
If you do not have any I/O, or even not enough I/O, the number of cores is completely meaningless. If the data isn't moving, it just determines how many cores are exploring the very bottom power saving modes, since there's nothing left to compute.
This is the sad, ugly truth that drove me from AMD's arms to Intel's, back in the Core 2 era. It doesn't matter how awesome your CPU is, if the data isn't moving. I really think Intel brought a whole new level of play with the ICHX south bridges, and they've been batting a thousand ever since, at least on I/O; That's the bedrock you need, to build a tower. Sand or mud, or even loose rock simply will not do.
NovusBogus wrote:Anyway, this sounds like it might be a motherboard bandwidth issue. I'm not aware of any useful quantitative studies narrowing it down but there's only so much data transfer a motherboard can handle before it starts to choke. I have the same issue at work when trying to do GUI type stuff while a couple of Xeon cores are pegged at 100%. Even terminal windows get a little bit laggy, which is no small feat.
ronch wrote:Hey guys, just noticed this earlier today. I was using CDBurnerXP to burn data on DVD and at the same time, formatting a USB flash drive (the long format, not quick format) as well as browsing the Net with Chrome. Now as per my Sig below you can see that I'm using an FX-8350 which means I can multitask like crazy. The thing is, the things I was doing weren't even supposed to begin to stress my system, and yet when I roll the scroll wheel on my mouse within Chrome there were short pauses before Chrome actually scrolls pages up or down. I just switched back to Chrome from Firefox because the latest version seems to be getting slow again so I'm not sure if this is just Chrome beginning to slow down as well, but I've always known Chrome to be snappy. Is this normal? Could writing to the USB have slowed things down? Or could it be the act of burning a DVD that's causing the short pauses within Chrome? Or did I just likely browse 'difficult' web pages? (Then again i usually surf just the same web sites all the time.) Thanks in advance.
Deanjo wrote:I noticed the same thing when utilizing USB storage on my 8350/990FX setup. Mouse got jumpy in Windows 8. Scoured the net and found a firmware update for the Asmedia USB 3 chipset after looking at the logs in linux and found tons of USB link renegotiations. After the firmware was updated, those disappeared.