Was:Replace my Geforce 8800 GT / Now: upgrade my whole PC

Hrm, nevermind. I might have misread a forum post somewhere. I'll have to look for more detailed benchmarks.

Thin_J wrote:

Hrm, nevermind. I might have misread a forum post somewhere. I'll have to look for more detailed benchmarks.

For the ~12 hours I played GW2 during the pre-launch weekend, I kept the resource monitor open on a second screen and my i7 2.9ghz cpu never went over 50%. For what it's worth, SR3 spikes the cpu to 100% constantly. Settings maxed on both games at 1920x1080 with sli'd gtx 670 4gb cards.

Tom's did a test of GW2 and like many other MMO's they found the CPU played a definite role in FPS.

ibdoomed wrote:

For the ~12 hours I played GW2 during the pre-launch weekend, I kept the resource monitor open on a second screen and my i7 2.9ghz cpu never went over 50%.

If you're not paying particular attention to the load it puts on each individual core then this doesn't actually mean anything. If it never went over 50% then I would be willing to bet you a whooooooooole lot that the game only makes use of two cores on your processor but was maxing out both those cores at 100%. Meanwhile you have two cores sitting idle doing nothing other than occasionally maybe doing something with windows background processes. So you sit at 50% or below the entire time the game is running.

The 560Ti runs just about everything at max settings, GW2 is a little much for my CPU i3-2120 3.4 ghz. But playing everything on high with just a couple settings on medium is cool with me.

Depending on your resolution, that might be the card choking, not the CPU. I haven't seen benchmarks on GW2 yet, so I'm not sure how CPU-dependent it is, but usually, if turning down graphic effects gets your frame rate where it needs to be, then it's the graphic card that's not keeping up, rather than the CPU.

Thin_J wrote:
ibdoomed wrote:

For the ~12 hours I played GW2 during the pre-launch weekend, I kept the resource monitor open on a second screen and my i7 2.9ghz cpu never went over 50%.

If you're not paying particular attention to the load it puts on each individual core then this doesn't actually mean anything. If it never went over 50% then I would be willing to bet you a whooooooooole lot that the game only makes use of two cores on your processor but was maxing out both those cores at 100%.

It was using ~50% of the 4 physical cores and little of the HTs. The resource monitor does break them down, here's an example from right now, obviously I'm not gaming. See how every other core is used very little.

IMAGE(http://i.imgur.com/OZNls.jpg)

Actually, I think you might be CPU-choked in that game, ibdoomed. If you'll notice, you've got at least 8 threads, maybe more (it goes off the bottom of the screen), all running at 50% of the pseudo-CPU they're on.

If you turn off hyperthreading, you'll get a much more accurate picture, and I suspect you'll see at least four cores, fully loaded, and maybe all six.

Yeah, I'd be surprised if Malor isn't right.

I'm not playing a game right now, when I just took that screenshot. And I'm not redownloading GW2 just to look at the cpu usage, it was a horrible game.

The benchmarks that have shown up are showing a far different picture than you're describing.

*Rephrased.

Thin_J wrote:

The benchmarks that have shown up are showing a far different picture than you're describing.

*Rephrased.

Ok, sorry. I will refrain from offering real world experience in the future.

And if I'm not going to do that, then what's the point of even visiting...

I'm just trying to understand how you can get such a wildly different result compared to people who test things like this all the time.

Oh, I misunderstood, ibdoomed, I thought your screenshot was from GW2.

I'd still rather see the results without hyperthreading. That really messes with measurements. For some things, it's a 15-20% overall throughput improvement, which is not small, but it makes it much harder to determine when you're actually at the limits of your CPU power.

My C2D is almost certainly choking my GW2 WvW performance (sub 10fps running around with a zerg on lowest settings), but I get a slight hint from the way developers in the forum have worded their replies that performance/utilisation isn't where they want it to be yet. GW2 seems a little temperamental in it's performance at times, and there's meant to be an nvidia bug where it goes out of performance mode and you need to alt-tab out and back to restore it.

I'd like to think the performance picture for GW2 will still change in the future, but I'm sure there's limits.

Malor wrote:

It's all really dependent on your income versus expense levels, AnimeJ. IIRC, you've got a big family, so probably the great majority of your income is already spoken for. Geeks tend to be single, and tech jobs tend to pay well, so they have lots of money for toys.

True enough, but the flipside of that is that I do tend to have large amounts that show up periodically; when I built my PC, I was sitting on quite a bit after spending a year deployed; I could easily have spent twice what I did and still have been fine. I'm just a really, really stingy bastard by nature with a strong aversion to paying in excess for smaller increments in improvement; that's really what kept me out of decent audio stuff for so long, more than the costs involved.

So today I managed to get the last part I needed for my new PC build in the mail. I've been trying to find a good balance of affordability with awesome. Spent the whole day assembling parts, figuring out where wires go and so on. I'm typing this now on my brand new built computer. Everything seems to be working great so far, though I haven't been able to install/play any games yet (Guild Wars 2 is currently at 20% downloaded, the Steam Client and its library of games will be next).

I wound up with:

Intel i5-3570k 3.40GHz (plus Cooler Master aftermarket cooling system to replace the stock fan)
Rosewill BLACKHAWK Gaming ATX Mid Tower Computer Case
ASRock Z77 Extreme 6 motherboard
8 Gig of RAM
OCZ Vertex 120 GB SSD
Seagate 1.5 TB HD
Windows 7 64-bit

and of course, the new items from my old PC I had bought when my graphics card died:
OCZ 700W power supply
and Geforce GTX 660 Ti

Cost was around $1100 spread over several weeks of purchases. As I said, it all seems to be working great so far. Only trouble I had so far is that 2 of the 4 USB ports on the back don't seem to work. I may try a BIOS update to the motherboard, or deal with it by use of USB hubs.

Can't wait for you to actually play some games. Should be good.

Only trouble I had so far is that 2 of the 4 USB ports on the back don't seem to work.

Those might be USB 3 ports, which usually do need a separate driver. ASUS is using a Renesas chip for its USB3 ports... dunno about ASRock.

That should be a pretty sweet setup.

edit: Yeah, I looked it up, and Newegg says you have two USB 2 and two USB 3 ports on the back panel, so the driver is almost certainly what you need to light them up.

Malor wrote:

edit: Yeah, I looked it up, and Newegg says you have two USB 2 and two USB 3 ports on the back panel, so the driver is almost certainly what you need to light them up.

Odd. All four ports in back are labeled clearly as being 3.0. Though it is odd they'd put two at the top and two at the bottom if there weren't some difference in them. I think you're onto something.

However, after a visit to the ASRock website to download both the Intel USB 3.0 drivers and the ASRock USB 3.0 drivers, those two ports at the top are still not working. Also all the USB 3.0 settings in the BIOS setup say "enabled" but no go. I'll keep picking at it.

But, I'm not going to let a little thing like that eat into my gaming time. So far, Guild Wars 2, Skyrim and Borderlands 2 have all been confirmed as running beautifully at the highest settings with this rig.

Well, you could have bad ports, that does happen sometimes. Might want to ask ASRock.

I also had an 8800 gts 300 meg version. When it died EVGA replaced it for free. That replacement died a few days later! So they upgraded it for free to a GTS 450 which I think is an upgrade but still in the same class. There is something I find really disappointing and lacking in every video card review article I have read including the ones at TomsHardware. They make recommendations on which card is best at various price points, but don't really make it clear to me what I get for moving from one class up to another. When I have tried to look into it I often find that the additional settings a higher class card make possible don't really add much to the visuals in.

OK, sorry, bit off topic, BUT RELATED

Tscott wrote:
Malor wrote:

edit: Yeah, I looked it up, and Newegg says you have two USB 2 and two USB 3 ports on the back panel, so the driver is almost certainly what you need to light them up.

Odd. All four ports in back are labeled clearly as being 3.0. Though it is odd they'd put two at the top and two at the bottom if there weren't some difference in them. I think you're onto something.

However, after a visit to the ASRock website to download both the Intel USB 3.0 drivers and the ASRock USB 3.0 drivers, those two ports at the top are still not working. Also all the USB 3.0 settings in the BIOS setup say "enabled" but no go. I'll keep picking at it.

I ended up with the exact same motherboard for my upgrade and ran into a similar problem. The four USB 3.0 ports don't seem to do so well when used with my keyboard and mouse, which includes the two ports labeled as intended for keyboard and mouse. The effect was that the keyboard and mouse wouldn't reliably power up after boot even though they'd work fine in the CMOS setup. What I found though was that if I used the USB 2.0 ports, everything worked fine. They're the two red ports in the middle. See the photos here for reference. I should note that this was all before installing drivers for my motherboard so the problem may have been fixed by that, but I haven't had any reason to go back and verify.

As for graphics cards... I think it's most useful just to see FPS results for games I care about rather than say whether a card is specifically better at FXAA or whatever, if that's what you're saying. Upgrading a PC is such a balancing act between the performance of different parts anyway that I don't know how useful a feature list would be. I think that beyond the point where a graphics card gives you a reliably good framerate at your desired detail settings (both FPS and detail being a matter of preference), getting a better card than that just provides a degree of future-proofing, which may or may not be more cost effective than simply getting a new graphics card at the point when the base choice doesn't meet expectations.

blueOrange wrote:

I also had an 8800 gts 300 meg version. When it died EVGA replaced it for free. That replacement died a few days later! So they upgraded it for free to a GTS 450 which I think is an upgrade but still in the same class. There is something I find really disappointing and lacking in every video card review article I have read including the ones at TomsHardware. They make recommendations on which card is best at various price points, but don't really make it clear to me what I get for moving from one class up to another. When I have tried to look into it I often find that the additional settings a higher class card make possible don't really add much to the visuals in.

OK, sorry, bit off topic, BUT RELATED :)

This might help. Doesn't give you all the nitty gritty bits and bobs differences like DX support or whatnot but lets you have a good idea whether it's worth upgrading or not. Usually you should go up two or more levels otherwise an upgrade generally isn't worth it - at least that's how it used to be.