I7 4790k. Will hyperthreading improve my live streaming.

Solution
Right, the i7 is clocked a little higher out of the box and I think that's what many benchmarks like cinebench etc rely on. Thing is the 4790k and 4690k are both unlocked and both capable of overclocking - when overclocked they reach essentially the same speeds, negating any speed advantage of the cores themselves.
i7's typically reach 4.7-4.8, i5's typically reach 4.6-4.7ghz.

They do have an extra 512k per core of l3 cache than the i5 which would probably work in tandem with the hyperthreading to keep the cores fed. At the base foundation though, they're the same exact cores as the i5. The i7's igpu is clocked at 1250mhz vs 1200 and while something is better than nothing I've never seen a major difference in such a little clock...
i would say i7 for streaming i5 for straght gaming i have both and i7 does much better for streaming get it i have 4690k and 4790k i7 is worth for streamers but for just gaming i5
 
Streaming doesn't only apply to what processor you use, yeah it might improve the quality but it also bases on your upload speed of your connection. Don't spend tons of money on PC parts for streaming if your internet connection isn't that great
 
tburdz, to see any difference between an I5 and ani7 streaming (on the same clock rate) you'd need to use extremely old, power hungry codecs.
Obs yields fairly performance light options which would have very little impact on framerate, so little, there wouldn't be a difference between the I7 and I5. And after all, it's 2015. Every half decent gpu released in the last 4 or 5 years supports gpu encoding. Which would result in ZERO difference between I5/I7.

I've played with hyperthreading back and forth and I must say there's absolutely no feelable (nor measurable using rounded fps averages) difference between in games or things related to gaming. Which might be of course related to having "only" one gtx 680.
What I see benefit in are rendering large files or building projects (visual studio).
 
But at least in my opinion for a intense streamer/content creator an i7 would prove helpful for someone who is not only streaming but following there stream multitasking possible rendering and an i7 would just not get outdated as quick as an i5. I think if u got the cash get the i7 but if u dont i5 is nothing to be ashamed of.

And dubbleclick i see were ur coming from but at the same time i just feel overall he would be better off with an i7
 
An i7 won't magically become better over time. It has it's advantages over the i5 now and those will stay. There won't suddenly come out more advantages. For a content creator, as in photo or video editing, programming (and building) and other tasks that put your system under continuous load, the i7 is worth it.
For real time execution and calculations like gaming, streaming or just casually doing stuff, there is no difference. Hyperthreading simply isn't a technology that helps in those cases. And you could have 500 programs running in the background while gaming, the i7 will still not yield anything over the i5, simply because how threading works. When a process is shown at 10% cpu load, that process doesn't really take 10% of the cpu's resources. It takes 100% at it's time, 0% on the other. A cpu does only know the state of working or idling.
 
not saying that hyper threading is why all hyperthreading is is a better way to schedule tasks a processor must complete but the extra cache and higher clock speeds help all im saying is if he got the cash get the i7
 
Is there any way someone who streams with an i5 or i7 can measure their performance? I don't see any direct comparisons and I'm sure it's frustrating for people seeking the answer. Something more solid than 'i7, it's better so get it'. Just like many people mistakenly say i7 is better in photo editing. Maybe on the rare occasion that calls for large (50-100 images in raw format at high resolutions above 2k x 2k) in batch processing. Otherwise, according to just about every photoshop benchmark out there the i7 has so little advantage it's not noticeable. For a $100 price premium over the i5 that's not good. Other content creation benchmarks follow suit and show the i5 right up with the i7. The only time I've seen any noticeable difference is in benchmarks comparing them out of the box and of course the i7 is clocked 500mhz faster out of the box. However they have almost the same top end when overclocked which means the i5 has more oc room than the i7 does.

Don't get me wrong, I know common thought (and selling points) are hyperthreading for multitasking but it would be much better to have some sort of evidence. Just because I 'feel' blue is a nicer color than yellow and many others agree doesn't make it so without some sort of meaningful data.

It's not that I'm somehow defending the i5 just because it's the choice I went with. I don't stream, have no desire to so am not affected by it. I'd just be curious to get some real data on the matter for those who do stream. If streaming follows suit like other multitasking or content creation comparisons between the i5 and i7 (both quad, not the higher end 6/8 core i7's) - then many people may very well be wasting a healthy chunk of money on an i7. I agree the quad i7 is better than the i5, $100 better not so much. Not from evidence I've seen and would rather someone put that money to better use or keep it in their pocket.
 
snphul i agree and i found some facts time will tell if they are true.

I will find more info asap and post here






Why i7 is "better"

Is hyperthreaded Yes vs No Somewhat common; Maximizes usage of each CPU core
Higher turbo clock speed 4.4 GHz vs 3.9 GHz Around 15% higher turbo clock speed
Higher clock speed 4 GHz vs 3.5 GHz Around 15% higher clock speed
Better PassMark score 11,249 vs 7,765 Around 45% better PassMark score
Better geekbench (64-bit) score 16,687 vs 12,880 Around 30% better geekbench (64-bit) score
Better PassMark (Overclocked) score 6,925.3 vs 4,652.1 Around 50% better PassMark (Overclocked) score
Better PassMark (Single core) score 2,534 vs 2,247 Around 15% better PassMark (Single core) score
More l3 cache per core 2 MB/core vs 1.5 MB/core Around 35% more l3 cache per core
Slightly better turbo clock speed 1,250 MHz vs 1,200 MHz Around 5% better turbo clock speed
Better cinebench r10 32Bit 1-core score 8,785 vs 7,619 More than 15% better cinebench r10 32Bit 1-core score
Better cinebench r10 32Bit score 33,538 vs 27,090 Around 25% better cinebench r10 32Bit score
 
Right, the i7 is clocked a little higher out of the box and I think that's what many benchmarks like cinebench etc rely on. Thing is the 4790k and 4690k are both unlocked and both capable of overclocking - when overclocked they reach essentially the same speeds, negating any speed advantage of the cores themselves.
i7's typically reach 4.7-4.8, i5's typically reach 4.6-4.7ghz.

They do have an extra 512k per core of l3 cache than the i5 which would probably work in tandem with the hyperthreading to keep the cores fed. At the base foundation though, they're the same exact cores as the i5. The i7's igpu is clocked at 1250mhz vs 1200 and while something is better than nothing I've never seen a major difference in such a little clock change for graphics. Regarding discrete gpu's, even bumping my hd 7850 from 800mhz to 925mhz made only a slight improvement. That's 125mhz vs 50mhz and would only apply to those using the igpu (most don't).

I agree they're performance increases, but very very minor ones for such a huge price leap. Nothing major like an extra core or two. The i5's run around $70 more than the i3's and give higher clocked cores, turbo boost and 2 additional cores compared to the i3's. The i7's add a couple mb's of cache and ht for $100?

I meant in terms of finding out a direct comparison between streaming with ht on or off (could likely be done with a core i7 by itself). If someone wanted to go through the trouble of testing their streaming software with their ht enabled and disabled to get a proper relevance of ht on it's own. An i7 with ht turned off would perform almost identical to an i5 and would give a clearer indication just how much benefit ht provides streaming in specific so people know whether it's worth the 50% price hike or not.
 
Solution
synphul i see what ur saying i agree i think intel should change it up a little and give i7 a bigger advantage. I suppose some of that price goes to saying that you have the best and i still think that at stock clock speeds they win but i think the next time i upgrade ill go extreme platform or i5. That said i still think the hyperthreading helps with video encoding which boosts rendering speeds but other then that i think i5 wins out for value upon the research i just did.
 
Just from a theory standpoint, my gripe is the pricing bracket since they placed the i7 quads in an awkward bracket.

i3 -> i5 $70 difference and gains 2 additional cores.

i5 -> quad i7 $70-100 and gains slightly higher stock speeds (negated if comparing k variants since they have nearly identical overclocking potential) and hyperthreading along with 2mb l3 cache.

quad i7 -> 6 core i7 5820k $30-40 difference from 4790k, loses a little clock speed, keeps ht and adds 2 more physical cores and goes from 8mb cache to 15mb of cache. Even though the platform is a bit more expensive, it offers a lot too - twice the supported system ram (to 64gb), almost 3 times the memory bandwidth and more pcie lanes. (the only thing I'm not certain of is clock speed potential overclocking a 5820k, I have a feeling they don't overclock quite as well as the quad cores due to the additional power consumption and people have reported some instabilities with overclocking on the x99 platform).

It's as if the biggest price leap between dual, quad and hex cores ranging from i3 to i5 to i7 is the one with the least performance bump. That's what makes me think it's awkward. Both physical improvements are the least costly price leaps (2-4 cores and 4-6 cores).
 
I see in some benchmarks that an i7 slightly beats out the i5 in gaming but thats the only advantage i see besides video encoding.

So my conclusion is that yes the i7 is better slightly but it is but is it worth it? I dont know

The pricing is weird like you said i dont get why the i7 is so much more or why it gets the i7 title i think it should be an i5-e for i5 extreme and then u have the i7 which is the 5820k then the extreme series would be the 5960x and maybe a 10 core cpu.

But i dont work for intel nor do i understand how it works over there so all i can do is say what i think.
 


I've been benchmarking with HT on and off fairly excessive. No streaming benchmarks, though, just recording (h.264 cuda, gtx 680) but that shouldn't make a difference. I didn't save my results, though. All I can remember is that building boost (fully library) and some personal projects yielded the highest profit with switching HT on. Sony vegas was barely noticeable (30s 1440p video, 60fps, 26k bitrate, cpu only) - on average between 2:07-2:13 without HT and 2:04-2:17 with HT. Rendering a 30 min 1080p video (60fps, 21k br) showed a small improvement, but I only did that test twice each. Times were 1:20 x2 without, 1:10 and 1:15 with HT, but I might mix something up here. Handbrake showed literally no improvement at all, either way 27 fps average.
I'll redo some guild wars 2 benchmarks right now.
 

the only reason the 4790k beats the i5 in gaming is because its clocked at 4ghz and Turbos to 4.4, while the i5 is at 3.5
 
I just did a quick benchmark of guild wars 2, always running the same path, NPC crowded area, 1 minute benchmarks of which roughly 20 second were still standing. Either way, recording with Shadowplay and Fraps, Benchmarking with Fraps. Running an 8 thread app I made at the same time: CreateRemoteThread as injection, being the first (main thread) which starts 7 other threads calculating more or less performance heavy stuff (one iterates through every agent ID on the same map (which gets up to ~400) roughly 10.000 times a second. For this benchmark I obviously restarted the computer before each and always started gw2, fraps, esp and then started recording and benchmarking.
Results can be found as download here: Benchmarks.rar.

PS: Use Excel to open them. Alternatively normal Text Editor works as well. The i5 even wins on average this time, but be told that doesn't really mean anything in such a little test pool. After all differences are negible. I may add some compilation benchmarks tomorrow, in case anyone is interested in how much benefit it does show in a programmers work lol.
 
That's awesome DubbleClick. I would have done it myself if I had an i7. I know it's really hard to say from one situation to the next since h.t. is program dependent (like you saw with handbrake). Hyperthreading can add a performance bump, but it's interesting to see in which scenarios and I wish large review sites would take the time to do this. I think it's really important to weed through the data in order to make a more accurate comparison.

At stock speeds sure a 4790k beats an i5 4690k because of the difference in clock speeds out of the box. I think comparing two 'k' series though levels the playing field a bit more since the i7 comes at 4ghz (4.4 turbo) and the 4690k i5 is at 3.5ghz (3.9 turbo). Overclock them (since comparing two k's and that's what k's are made for) and the 4790k ends up around 4.7 to 4.8 tops. The 4690k reaches 4.6 to 4.7 tops in most cases - and with both i5/i7 some overclock a little worse, some a little better (golden chips). The i5 covers a lot of ground when overclocked and pretty much closes the gap in core speed.