Uppgraded pc worse fps (wow)

[Barnabas]

And how much did your Xeon/GPU cost compared to my i5 (280 euro) + GTX 1070 (520 euro) which plays on 10 with max shadows and custom AA at 1440 (downscaled to 1080) at 70-120 FPS?

$50. Read it and weep. Sure and that's not in raids I bet lol. You get the same 30fps everyone else gets haha. Mine isn't downscaled either.

[Lathais]

Nonsense. Wow uses 3-4 threads since WoD. Clock speed is still very important, but try playing it on a dual core (GLHF).

It uses many more threads than that, however, the primary thread is still on large thread and how well that thread can run largely determines your performance. Also, plenty of people play on Pentium G3258s just fine.

- - - Updated - - -

Or just shut off your additional cores and see what happens to your FPS. WoW may not be a terribly well designed game for the use of multi-cores (maybe that has something to do with blizzard insisting the game to be playable on low-tier hardware), but just because it isn't draining all it can from the additional cores doesn't mean you're not gaining significant performance from them.

http://www.tomshardware.com/reviews/...e,3849-10.html

I don't see a signifigant performance gain going from the OCed G3258 to the i5. Virtually no difference between the OCed G3258(Dual Core, No HT), the i3(Dual core with HT) and the i5. HT and additional cores do very very little for WoW. Based on seeing the performance gain when going from the Athlons, which have a much lower IPC than even the Pentiums, you see it is indeed clock speed and not core count that matters. The rather sizable increase from OCing the G3258 shows this as well.

[Barnabas]

There is your answer, 2Ghz less

Max speed is 3.2ghz actually

[Lathais]

Max speed is 3.2ghz actually

Yet his old CPU turboed up to 3.8 all on it's own and he had it OCed to 4.5, which you just can't do with the Xeon. No matter how you look at it, his old CPU was better for WoW.

[Barnabas]

Yet his old CPU turboed up to 3.8 all on it's own and he had it OCed to 4.5, which you just can't do with the Xeon. No matter how you look at it, his old CPU was better for WoW.

Very little difference is closer to the truth.

[Adam552]

Use the old CPU, you bought a Xeon which is really for workstations/servers and not for gaming.

[Hottage]

$50. Read it and weep. Sure and that's not in raids I bet lol. You get the same 30fps everyone else gets haha. Mine isn't downscaled either.

Your CPU and GPU cost $50 together? I doubt that.

I get 70-120 FPS everywhere, easily. Neither my CPU or GPU go above 70 degrees.

Also, it's rendered at 1440 and downscaled because my monitor only supports 1920x1080.
IE: Downscaling is a good thing here, it gives a much crisper image.

Not sure what I'm supposed to cry about :-/

[Gorgodeus]

Very little difference is closer to the truth.

It is a huge difference. That Xeon CPU does not even rank in the top 100 on single core performance. Think about that for a moment.

[Lathais]

Very little difference is closer to the truth.

Not really. The difference is pretty big, as evidenced by the OPs huge loss, even with a much better graphics card.

[Deleted]

http://www.tomshardware.com/reviews/...e,3849-10.html

I don't see a signifigant performance gain going from the OCed G3258 to the i5. Virtually no difference between the OCed G3258(Dual Core, No HT), the i3(Dual core with HT) and the i5. HT and additional cores do very very little for WoW. Based on seeing the performance gain when going from the Athlons, which have a much lower IPC than even the Pentiums, you see it is indeed clock speed and not core count that matters. The rather sizable increase from OCing the G3258 shows this as well.

Virtual cores and actual cores are not the same at all. Hyperthreading exists for CPU intensive tasks like bioinformatics and rendering videos, most games (including WoW) get minimal to no benefit from it. This says nothing about how much benefit a game can get from multiple physical cores.

Also, you can't really compare CPUs like that. Superficially they may seem to have the same stats, but under the hood there are almost always major differences. Notably multi-core CPUs have a "dominant core" that's more powerful than the others, but you have no way to tell how much more powerful without testing it.

That's why I suggest that you just turn off the cores on your own CPUs to test the performance difference. That is the only clean way. Of course you won't see massive improvements on a multi-core setup (again, that's most likely intentional design), and of course you see a big change in performance from upping your clockspeed, but there still is an easily measurable difference in using one vs several cores.

[Lathais]

Virtual cores and actual cores are not the same at all. Hyperthreading exists for CPU intensive tasks like bioinformatics and rendering videos, most games (including WoW) get minimal to no benefit from it. This says nothing about how much benefit a game can get from multiple physical cores.

Also, you can't really compare CPUs like that. Superficially they may seem to have the same stats, but under the hood there are almost always major differences. Notably multi-core CPUs have a "dominant core" that's more powerful than the others, but you have no way to tell how much more powerful without testing it.

That's why I suggest that you just turn off the cores on your own CPUs to test the performance difference. That is the only clean way. Of course you won't see massive improvements on a multi-core setup (again, that's most likely intentional design), and of course you see a big change in performance from upping your clockspeed, but there still is an easily measurable difference in using one vs several cores.

Throw the virtual cores out then. Look at the performance of the 4690k and the G3258, which are basically the same, just the i5 has 4 cores. They are both Haswell CPUs using pretty much the exact same technology. The OCed G3258 is running at 4.5gHz, the 4690k at it's turbo boost of 3.8. You can see that yes, there is in fact a difference, because it is taking higher clock speeds on the G3258 to achieve near the same results, actually slightly under. However, You can also see, from the AMD CPUs listed, that IPC is much much more important. It performs worse than the G3258 without an OC. More cores matter slightly, mainly because yes, WoW is multithreaded, so if you can leave the one main thread on a core by itself, it runs faster. Having multiple cores to move all the minor shit too does in fact help. But not as much as IPC. IPC is king to WoW. The IPC of the Xeons is far worse than that of even the G3258 and especially the i5's. Xeons are not a great choice for WoW for that reason.

[Barnabas]

Not really. The difference is pretty big, as evidenced by the OPs huge loss, even with a much better graphics card.

That's because OP didn't set affinity mask inside the game client. It's the same as the old chip when you do that.

[Bigvizz]

That's because OP didn't set affinity mask inside the game client. It's the same as the old chip when you do that.

I believe the "Set Affinity Mask" thing stop being a thing a long time ago. At least for using more cores, it's really only useful for assigning specific cores.

http://wowwiki.wikia.com/wiki/CVar_processAffinityMask

[Deleted]

Stop spreading misinformation, WoW benefits from quad cores. Try playing it on a Skylake i3 compared to a Skylake i5 at the same clock speeds, the I5 will always win, though obviously you haven't bothered to test this yourself.

None of which changes the fact that the graphics portion is handled primarily by 1 core, and the single core performance of a CPU is more important than number of cores, which is what this thread is about. The other cores are used primarily for menial tasks in-game.

The i5 having 50% more bandwidth has more to do with the difference than anything.

If number of cores were the main thing, the OPs CPU would romp. It is more about single core performance.

Out of curiosity I did a short test with an i5-6600k in Legion beta. Everything was set on stock speed, with turbo turned OFF, first time I disabled 2 cores, then enabled all. Path: mage order hall ->Dalaran "entrance" -> center -> flight point -> Azshara. Though not fully crowded, a decently populated Dalaran is on of the places that may present a challenge for a cpu currently, excluding badly configured raid encounters, bugged zones, outdated addons etc. Everything went smooth with 2 cores, hardly any difference, once I caught a slight drop of fps, that means I saw 5x fps instead of 60 for a second or so. GPU was the same GTX 950 (an old trash compared to the OP's 1070), all ingame settings the same (generally the game's auto/recommended level for this config, mostly high with level 5 view distance).

Now, a few minutes long test without exact measures is not evidence for anything. Still, it is easy to see that any outdoor scenario, questing, flying or whatever else that is generally less taxing on the hardware than the capital city should produce similar results. Until I see massive fps drops or stuttering in a raid with 2 cores, I fully agree with Gorgodeus.

[Bryntrollian]

My xeon is really fast. I play at 7 in 1440p and invasions are the only thing that slows it down. 30+ fps is playable though.

Not nearly as fast as his 4670k though, which is all that matters...

[satimy]

I have a 4790k and a 1060 and j get massive fps drops

[Deleted]

Very little difference is closer to the truth.

Well, no. Loads better - since WoW's thread load will look something akin to this in a typical raid environment. Note that this is on a i5-3570, the highlighted thread is at the limit of what the CPU can handle, and is limiting the game's performance.


While there are obvious benefits to adding a second core here, the diminishing returns after that are high. A third or fourth core can still reduce thread service delay, but it won't translate into a significant FPS gain.

Note that in cities, due to the considerable amount of asset loading required - thread distribution can change a bit to look more like this:


This never lasts for long however, and is not something you should build a machine around.

[simonlvschal]

The graphics portion is handled by 1 core, and is why 1 core will max out while the others are barely used. Stop spreading misinformation.

- - - Updated - - -

To the OP:

The single core performance of your new chip is pathetic.

Have fun scrolling down to find it:
https://www.cpubenchmark.net/singleThread.html

It only scores 1709, to help you find it easier.

stop spreading lies dude.. try look at your cpu again.. i have a i7 5830k and mine uses 3 cores in wow where 1st core is 100%, 2nd around 70-90 and 3rd around 60ish. the only reason for why 1st core is 100% is due to watching streams and other things that uses 1 thread...

[Gorgodeus]

stop spreading lies dude.. try look at your cpu again.. i have a i7 5830k and mine uses 3 cores in wow where 1st core is 100%, 2nd around 70-90 and 3rd around 60ish. the only reason for why 1st core is 100% is due to watching streams and other things that uses 1 thread...

And another one chimes in that does not have a clue.

[simonlvschal]

And another one chimes in that does not have a clue.

sorry but i speak the truth.. learn to use tools your Pc comes with when getting wins 10.. or go and fetch softwares that gives u a idea.. its really easy to find out how much your cpu use. just check your Joblist easy as pie. and sorry you feel so bad when people got better pcs then u.

oh please dont talk like you know something when you dont..

come again when u have a eduction in computer science

or something that gives you propper knowledge in how a cpu works and how programming works