RTB

It really depends on what you're going to do with the new rig. There aren't many programs that can take advantage of hyperthreading on the 2600(k), so you can choose to downgrade to a 2500(k) and save quite a bit of money. The multiplier unlocked version of the i7 2600 (aptly named i7 2600k) is just a teensy bit more expensive, and it'll let you run 4ghz easily if you care for it. Likewise, I can't recommend a regular 2500. 4 DIMMs for 8GB makes no sense, that should be 2x4GB instead. While it would cost another 100 dollar or so, a 64GB SSD would be a great addition. You can either go for the Crucial M4 or Intel 320/510 (as well as a couple others) if you want to avoid potential issues with sandforce drives. And finally... I think you can build your own rig for 900 dollar and have it match whatever Dell is offering.

I should try a test at my university, though they're still on firefox 3.6 and exploder 7 or so. The majority of official student housing here has 100mbit symmetrical available per student, and of course mine doesn't.

It's the typical geek solution. Having a small issue with a single program? RECOMPILE THE WHOLE KERNEL, TWICE.

I do wonder if this is a thorough way to test ping. Most of the time I'm wondering if the connection here is stable, I hop onto a local quake3 or quakelive server and check the lagometer. Since those packets are being sent anywhere between 40 and 125 times in a second, you can really see any little stutter out there. It would be cool if CA3LE developed a method to test ping stability at low bandwidth and at high bandwidth usage (aka during a speed test), but with similar granularity. Since TCP is so sensitive to ping, this would explain most of the drops and thus lower speed results if worked out properly,

Conveniently leaving out USB3 and eSATA in the graph. Being as generic as USB it does seem good, but I'm not sure how licensing is gonna work.

I felt 2010 for sure. I think I'm about 2011 now.

*facepalms*

How do you think it would operate then? The only way to make a CPU run cooler with software is to drop the speed and/or the voltage. Cool'n'quiet does this depending on CPU usage, so ideally if you need the whole 3.6Ghz of processing power the CPU would instantly (well... delay measured in microseconds) work at full speed.

Well that's odd indeed. The hop where it goes wrong is present on both routes. I wonder if it's the same trick Comcast is doing... which is to have a single 10gbit line to the outside world, which is not nearly enough, so unless content providers put servers inside Comcast's network Comcast customers can barely reach their website. Evil, but unavoidable in this max-profit world.

I'm with dlewis23 here. There's no explicit split in half thing about bandwidth, the only rule is that the throughoutput of both (or more) connected devices combined cannot surpass whatever limit the object they're connected to has. The bandwidth graphs he showed have one correlation, the combined bandwidth is 100 mbps or less, because that can't be surpassed. In a case where the combined bandwidth would want to go above 100 mbps, QOS decides which server/service gets priority. Makes perfect sense to use it in the case of multiple servers on a single link. As for the 54mbps thing of wireless... that's highly theoretical. Hubs are simple devices, they do not check for collisions at all. Unless you want the traffic to be cloned, they are to be avoided.

The odds of unlocking an Athlon X2 are bad, because almost all are native dual cores. If you want a chance at unlocking try an Athlon X3, as those are genuine quad cores with a possible dead fourth core. That PSU should be able to handle that rig, but Rosewill isn't known for quality PSUs. You can find faster memory for that price, such as this.

1) My guess is a faulty temperature sensor for core 3. Does it crash if you put any CPU stress programs on just that core? 2) First thought is driver issue. If there's a newer driver out there, try removing the old one completely, then install the new one. On second thought, you're running an odd amount of memory. This puts strain on the memory controller, so if the CPU is bad, that certainly isn't going to help. Your video card itself is one known to run hot, might be a faulty application of thermal paste causing it to overheat. Just saying 'I have an 850 watt PSU costing 400 dollar' doesn't say much btw. Brand name does matter because it's easy to rate a PSU too high and then pretend it's high-end by giving it a high price. As a sidenote, more wattage isn't always better. A PSU runs the most efficient in the 20-80% area of its capability, so if you use one capable of 850 watt when your system is around 90-200 you're needlessly converting electricity to heat.

If you're on DSL, you should know that the linespeed is very dependent on the distance from the hub you're connected to. This isn't something your ISP can do much about, but it's a positive thing for them anyway. It's not something they should keep a secret though, that's just lame.

On the other hand, if you have reasons to assume the connection to your own ISP is malfunctioning, a download test from within their own network is better because in that case you'll be sure any problem must be anywhere between your computer and your ISP. Which ironically is a reason for an ISP to never offer such a download test, since they can't blame the internets then. Though what I've experienced here is that the helldesk blames the router repeatedly 'till the caller gives up.

Then I guess my source was wrong. But still, that does make me wonder why there's so little (if any) being done to upgrade those underpowered intercontinental connections. If the government prefers to spend money on local wiring, that kinda implies they prefer to keep Australia mostly secluded from the rest of the interwebs.

To the best of my knowledge internet access in Australia tends to be monopolistic aka silly bandwidth caps (think < 10 GB for a 8/1 connection or so), or hardly available at all if you're not in the most densely populated regions. If there was something sane that could push rules/fines on ISPs an accurate testing site would be very useful. Cynical me says there isn't one. Also CA3LE, Germany and Holland are very central to Europe and have the server connections you need for sure.

Right now I'd fit in a group called 'Lazy Moderators' then. Real life has priority over virtual life. That's how it is and should be.

UK? That's one wrong guess And that comment was basically from the days I had access to a 8/1 connection which I could max out to here. Been ages since I last tweaked that stuff too. As for the settings, it was VanBurens stuff, along with some random calculator for the exact rwin depending on maximum theoretical download speed and avg ping to download sites. Oh look, I'm still an inactive moderator while active on the forum. I guess CA3LE noticed my lazyness.

In related news, the 1 GB download test is coming. Overkill, but that's just how CA3LE like it. I think you can max out no matter the distance, it just takes tweaking of the receive window, and maybe some other values.

You're certainly going to notice a difference going from single to dual core, a single program will have more difficulties hogging the computer to itself, so it ends up being much more responsive. The other thing that would make a large difference is an SSD instead of a plain hard disk, but those aren't affordable yet for low-end budgets.

I thought you were going for this? Anyway, that's a budget dual core by Intel. It'll do, but compared to AMD's offerings in that price range they are normally inferior. Intel's integrated graphics 'card' is pure crap for gaming, it is made for cheap 2d displaying and nothing else.

the "Core 2 Quad Q9450 2.66GHz w/4 gig ram" will be the fastest, though if you are an excessive multitasker you may see a benefit from using 6 GB. I wasn't aware anyone ever sold a 50 GB drive, so it will have to be old and be the bottleneck for your system. That video card you mention is the absolute slowest discrete card nvidia sells at the moment, so don't expect any performance from that one either.

Let's take a more simple approach: You're all spammers!

On the other hand, he would have no time to post in that case... tempting.

I'll just repeat my post from the other thread here: I also vote for changing shug's title to "Karmaddict".