|
On December 25 2011 12:16 BBallTime wrote: hi guys i would like to know how the 2600k fares vs the 2500k in streaming...i would like to stream in 1920x1200 or 1920x1080!
On December 25 2011 20:36 BBallTime wrote: Since i also dont have a really great upload...
First, let's establish what your actual upload is? Without the required upload, either cpu won't help you. You need good upload for what you're trying to do. Preferably 4+ Mbps.
On December 25 2011 20:36 BBallTime wrote: i read somewhere that xsplit cant make use of hyperthreading but i think thats a few months old so has it changed now? with its hyperthreading, how much better is the 2600k if i optimise right? You probably read that XSplit had an issue using a multi-core CPU (not hypertheading). This has been fixed 6+ months ago. There really is no way of telling you how much better it will get. How would you measure it? The best answer I can give, is that you will be able to do more with the stronger cpu.
Is a 2600k needed for streaming 1080p Starcraft 2? No. Will your stream benefit from a 2600k? Yes.
|
United Kingdom20157 Posts
On December 25 2011 20:36 BBallTime wrote: hey Cyro thanks for all the info,
i read somewhere that xsplit cant make use of hyperthreading but i think thats a few months old so has it changed now?
with its hyperthreading, how much better is the 2600k if i optimise right? Since i also dont have a really great upload...
The difference is about 20%.
I think i can be blamed for mentioning xsplit wouldnt utilise hyperthreading, i had issues with it for quite a while (that still exist sometimes, seemingly at random) and made an assumption i shouldnt have, since then ive had it working across 6-8 threads just fine
|
On December 25 2011 21:09 Cyro wrote: The difference is about 20%.
While this is true from a view of pure processing power, we have no evidence that this will transform or scale into streaming results with any known streaming program. XSplit is known to be having issues scaling in general.
|
hmmm, how well will the 2600k do with its hyperthreading vs the ivy bridge 2500k equivalent without hyperthreading?
cuz theres going to be a deal for the 2600k soon and i dont know if i should get it or just wait out for ivy bridge 2500k equivalent.
|
United Kingdom20157 Posts
On December 25 2011 21:16 HellGreen wrote:While this is true from a view of pure processing power, we have no evidence that this will transform or scale into streaming results with any known streaming program. XSplit is known to be having issues scaling in general.
I saw an x264 encoding benchmark earlier, the 20% is after the nonlinear scaling of adding extra threads. I tried to find it just now, but seems to have dissapeared... I think it is right, that is around the gain i saw from enabling HT on my set up (allowing me to drop a preset and increase fps from 20 to 24 without maxing cpu)
|
United Kingdom20157 Posts
On December 25 2011 21:30 BBallTime wrote: hmmm, how well will the 2600k do with its hyperthreading vs the ivy bridge 2500k equivalent without hyperthreading?
cuz theres going to be a deal for the 2600k soon and i dont know if i should get it or just wait out for ivy bridge 2500k equivalent.
Im guessing they will be about equal, we dont really have concrete evidence of ivy bridge performance or overclocking capabilities yet, not being released til march/april anyway i think.
|
On December 25 2011 21:33 Cyro wrote: I saw an x264 encoding benchmark earlier, the 20% is after the nonlinear scaling of adding extra threads.[...]
Yeah, but there's more going on than x264 here, that's why I wrote Xsplit and not x264 . Frame grabbing is one of XSplit's weaknesses right now, and I don't know if throwing extra power at it will help atm.
For what it is worth though, here's a x264 benchmark: http://www.techarp.com/showarticle.aspx?artno=669&pgno=3 On page 6, I find it particularly interesting that the i5 2500k @4.5 and @5.0 are yielding the same results.
|
United Kingdom20157 Posts
On December 25 2011 21:54 HellGreen wrote:Show nested quote +On December 25 2011 21:33 Cyro wrote: I saw an x264 encoding benchmark earlier, the 20% is after the nonlinear scaling of adding extra threads.[...] Yeah, but there's more going on than x264 here, that's why I wrote Xsplit and not x264 . Frame grabbing is one of XSplit's weaknesses right now, and I don't know if throwing extra power at it will help atm. For what it is worth though, here's a x264 benchmark: http://www.techarp.com/showarticle.aspx?artno=669&pgno=3
At 1280x720 frame grabbing has no performance hit whatsoever for me, and it is relatively minimal at maxed settings, the weakness isnt in massive CPU usage, it just seems to hit game FPS with the actual stream, encoding and CPU usage being unaffected. It does behave quite weirdly but i think your issues (hellgreen) are more with weaker system being hit harder somehow, a decently overclocked 2600k will easily more than double the performance of a stock 920. I dont know WHY the framerate hit happens with screen capture so i cant really comment other than that, but it hits your system significantly harder than mine for some reason
|
United Kingdom20157 Posts
On December 25 2011 21:54 HellGreen wrote:Show nested quote +On December 25 2011 21:33 Cyro wrote: I saw an x264 encoding benchmark earlier, the 20% is after the nonlinear scaling of adding extra threads.[...] Yeah, but there's more going on than x264 here, that's why I wrote Xsplit and not x264 . Frame grabbing is one of XSplit's weaknesses right now, and I don't know if throwing extra power at it will help atm. For what it is worth though, here's a x264 benchmark: http://www.techarp.com/showarticle.aspx?artno=669&pgno=3On page 6, I find it particularly interesting that the i5 2500k @4.5 and @5.0 are yielding the same results.
The benchmark i saw showed more of a gap with and without HT, and near linear scaling with clock speed to FPS output, its odd that yours shows the same results @4.5ghz and 5ghz, not sure why that would happen
|
i heard ivy bridge wont be much of an increase in performance compared to sandy bridge so i might just go with the 2600k hmm
|
On December 25 2011 23:17 BBallTime wrote: i heard ivy bridge wont be much of an increase in performance compared to sandy bridge so i might just go with the 2600k hmm Well, clock-for-clock, there should be about the exact same improvement as when Sandy Bridge got introduced. 17% (Sandy Bridge > Lynnfield) Source 17% (Ivy Bridge > Sandy Bridge) Source
As for performance in streaming (Xsplit), I'll repeat my statement and say: We will have to wait and see. Nothing suggests that it will scale, but I do make the caution that XSplit is still under development.
|
Just my two cents:
If you're that serious into streaming, you wouldn't want to spare any expense to get the best possible stream. And you wouldn't worry about sinking money into a 2600k only to upgrade to something better in a few months. It comes with the territory.
Otherwise, just get a 2500k.
|
Wont matter at all.
If you dont want to waste money, buy the 2500k. It is that simple.
|
Yeah, a 2600k is a total waste of money for people who make a living doing streams.
|
On December 26 2011 05:09 creepcolony wrote: Wont matter at all.
If you dont want to waste money, buy the 2500k. It is that simple.
An average increase of 20% additional performance for average use, and up to 30% for certain kinds of use (encoding being a big one) "doesn't matter," especially when we're talking about streaming that may or may not involve teh moniez?
Right.
2600k is easily recommendable for people who can get good deals and plan on utilizing it to the fullest extent. For normal gamers, 2500k would be enough, but we're not talking about just games like SC2 that are terrible at utilizing multiple threads; we're talking about encoding.
However, OP, if your upload isn't even 4Mbps or higher, you wouldn't get too much out of a 2600k over a 2500k unless you can get a really good deal. Also, the performance increase isn't necessarily something you'll want to go head over heels about; you need to weigh if the additional cost will pay for itself in the end.
|
United Kingdom20157 Posts
On December 26 2011 06:44 Zeke50100 wrote:Show nested quote +On December 26 2011 05:09 creepcolony wrote: Wont matter at all.
If you dont want to waste money, buy the 2500k. It is that simple. An average increase of 20% additional performance for average use, and up to 30% for certain kinds of use (encoding being a big one) "doesn't matter," especially when we're talking about streaming that may or may not involve teh moniez? Right. 2600k is easily recommendable for people who can get good deals and plan on utilizing it to the fullest extent. For normal gamers, 2500k would be enough, but we're not talking about just games like SC2 that are terrible at utilizing multiple threads; we're talking about encoding. However, OP, if your upload isn't even 4Mbps or higher, you wouldn't get too much out of a 2600k over a 2500k unless you can get a really good deal. Also, the performance increase isn't necessarily something you'll want to go head over heels about; you need to weigh if the additional cost will pay for itself in the end.
On the contrary, the difference is more important on lower bandwidth i think as you can compress more into your limited datastream
|
Anyone have any idea if the next gen 6xx series nvidia video cards or i think they're called gk1xx can hit the pci express 2.0 limit and actually put pci express 3.0 into use?
Can ivybridge overclock more than sandy bridge?
|
5930 Posts
On December 26 2011 15:22 BBallTime wrote: Anyone have any idea if the next gen 6xx series nvidia video cards or i think they're called gk1xx can hit the pci express 2.0 limit and actually put pci express 3.0 into use?
Can ivybridge overclock more than sandy bridge?
99% no. Possibly but I wouldn't count on it.
|
Ivybridge is a die shrink and K variants are expected to have a max multiplier of 63 so yes they're going to be better overclockers.
You have extremely high hopes if you think Kepler will saturate PCI-E 2.0.
|
On December 26 2011 15:29 skyR wrote: Ivybridge is a die shrink and K variants are expected to have a max multiplier of 63 so yes they're going to be better overclockers.
You have extremely high hopes if you think Kepler will saturate PCI-E 2.0.
Cooler temps, higher overclocks. Glad I'm waiting for ivybridge.
|
|
|
|