The improved performance probably comes mostly from cache hit rate as no longer the threads push each others data out of cache.
@imnotusingmyrealname456611 ай бұрын
This makes a lot of sense. We have seen how important cache is for CPU performance.
@Gibus212507 ай бұрын
With HT, when 2 thread works on the same data AND run on the same core (from the the same processus), it can increase a lot the performance. But as a developper view, wa cannot contains our thread into a specific core, it s the OS who choose, and yes, if 2 thread of different processus run on the same core, they share the L1 data cache, and increase the cache miss prpbability!
@NrGSourcEsurfeR6 ай бұрын
I9 i7 shouldnt really rock kuch differnt depending how you set them up, hyperthreading depends on the game your playing because not all devs are great at implementing it
@silfrido17686 ай бұрын
And people say L2 cache per core doesn’t matter…. I swear people say anything & think they are correct. I’m not saying you btw lol
@LBXZero11 ай бұрын
For hyperthreading/SMT in the modern architecture, each CPU core has at least 3 pipelines per core and several ALUs and 1 or maybe 2 FPUs. The primary pipeline has a fetch/decode unit. The secondary pipeline in an hyperthreaded/SMT core also has a fetch/decode unit to serve as a second thread. The tertiary pipelines are given hardware accelerated operations like superscalar operations. All of these pipelines share the ALUs and FPU(s) of the core. Given the low FPU count versus ALUs, games that utilize heavy FPU performance on the CPU may benefit from disabling hyperthreading/SMT to relieve the demand on the 1 or 2 FPUs available. The point of hyperthreading was making use of the secondary pipeline that executed the MMX, SSE, and other hardware acceleration instructions when such operations were not being called. The main pipeline will have data set up for a hardware accelerated function and when executed, the CPU firmware will issue a series of instructions (no fetch/decode stages) through an available secondary pipeline and do such data processing using the extra ALUs. Hyperthreading was about putting that secondary pipeline and extra ALUs to work when the applications don't use the related instruction libraries. Your performance improvements can come from adverse effects from sharing the various ALUs and few FPUs. Another improvement could be less thermal throttling from running less parts per core, allowing higher clocks. Many possibilities that need a bit of studying to figure out.
@PAIN-ot4cj11 ай бұрын
Great reviews, these are the type of scenarios I search for on google or reddit. Glad you made these videos.
@Code-n-FlameАй бұрын
I fully crippled my 12900k. E cores and HT disabled and OC'd the p cores to 5.4 GHz lol CBR23 points dropped from 27k to 15k but Hogwards Legacy now runs smooth as butter. The more you know.
@XiaOmegaX11 ай бұрын
With HT on, every time the core changes process context it has to dump L1 and L2 cache, do thing #2, dump cache again, do thing 1# again, repeat over and over.
@randomsaloom72385 ай бұрын
now we need a comparison between these results and having e-cores disabled
@malceum11 ай бұрын
Very good test, but I would have liked to see BF2042. It's by far the most CPU demanding game around. It uses 17 threads if the CPU supports it, and I'm curious how the game would perform with most of those threads being put on e-cores with HT off. Historically, Battlefield games have benefited from hyper-threading.
@lon3wolf64211 ай бұрын
Add Cyberpunk 2077 as well, turn off SMT support and you are in a world of pain.
@crazycop777411 ай бұрын
it stutters like hell. That's what happens when you turn off HT on bf2042.
@bamboo966610 ай бұрын
Cyberpunk use all of my 20 threads..
@cameron481a9 ай бұрын
@@crazycop7774 I actually didn’t find this to be true for me. Bumped my fps and stable in bf2042
@facegamefps67259 ай бұрын
@@cameron481asame here 6.0 4.6 8400 ddr5
@injurera2 ай бұрын
A suggestion. Disabling HT may help to achieve slightly higher OC ratios. E-cores are grouped in groups of 4, each sharing the same L2 cache. Leaving 1 e-core per group will make it having access to more cache, and allows you to have additional thermal headroom for OC. So 8 pcores plus 4 ecores should be the most optimal for gaming as they rarely use so much threads. I would be definitely interested in testing if such a strategy in fact gives perf benefits. Note that diabling all ecores is bad, as the windows scheduler goes weird, resulting in some performance anomalies.
@leandrodrace11 ай бұрын
Nicely done, you just forgot to mention lower temperature and power consuption when HT is off. I`ve been using HT/SMT off for years now just for that temp difference and slightly lower Vcore required in order to run my CPUs 100 to 200 MHz higher... had a 4930k before and now R9 5900X both very different but gave me the same headroom for overclock when doing this trick...
@pamus624211 ай бұрын
WOW! So simple yet so practical.
@dakrawnik420811 ай бұрын
lol try UV'ing instead to lower power draw. you're welcome.
@jakedill130411 ай бұрын
That's strange I was able to get a 4930k to a very stable 4.4 just by increasing the multiplier a little bit.. and it never really needs it except for spikes.. and this is with hyper threading on and a cooler that was about 30 bucks when I bought it in 2013... And I haven't pasted it since 2013 LOL. And I thought I lost the silicone lottery cuz I couldn't hit 4.6...
@Kryptic104611 ай бұрын
@@dakrawnik4208 - If you're overclocking a CPU, it requires whatever voltage it requires to be stable at that frequency. Turning HT off means the CPU requires less voltage to be stable at the same frequencies. Undervolting is really only a thing if you're running at a stock configuration. Once you overclock past a certain point, you will almost always need more voltage to run stable, not less, so I'm a bit confused as to why you would tell someone to undervolt an overclock. That's usually a recipe for instability.
@leandrodrace11 ай бұрын
@dakrawnik4208 well considering when stock it hits past 1.5 vcore I'm already on underlvolt with 1.337 vcore for 4950 on the best ccd and 4750 on the worse ccd. If I had SMT on that static 4950 MHz would be impossible. But on the old i7 extreme I needed 1.44 vcore to run 4.8GHz, like the other guy said after you past a point on the V/F curve it's not possible to undervolt...
@JamesSmith-sw3nk11 ай бұрын
I have a 13700k pc for my office and a 5800x3d for my gaming pc. I REALLY like how smooth my 13700k pc is. I can cpu mine with the "P-cores" and still use the browser, watch 4k youtube video, use social media, even play older games without any stuttering on the "E-cores". I think my next gaming pc will be Intel.
@jordypolfliet69115 ай бұрын
Your mom or boss does not understand why the energy bill rises even while just installed 20 solar panels. Paying 4000 USD/year for your mining expedition which gives you the equivalent of 200-500USD. CPU mining needs to be punishable by law.
@cemsengul164 ай бұрын
Yes there needs to be punishment by law.@@jordypolfliet6911
@JasonChen9274 ай бұрын
Yeah as long as there is no power usage concern, multi-tasking with E-core is very efficient.
@scoso13132 ай бұрын
I was blown away by the performance of the Raptor Lakes. Very big difference multitasking compared to my old 8700k. Haven’t used an AMD chip for about 20 years now but they look great for gaming now. I just wouldn’t want to lose the multitasking performance Intel currently offers.
@SkilliardАй бұрын
I'm assuming these tests were down with very few other processes running for the sake of consistency, which also makes it an unrealistic scenario. The average gamer is going to have a lot of other processes running- Discord, Steam, maybe a web browser, maybe a music app, etc. Hyperthreading works best when there's more of a load. Also worth considering that over time games get more heavily threaded, HT might become more useful. I remember people used to have the same conversation about the i7 6700k, saying you don''t need HT because you can just OC a 6600k. Well after a few years the 6600k would stutter while the 6700k did fine due to HT
@imnotusingmyrealname456611 ай бұрын
In Escape from Tarkov there's an option that's on by default that also only uses the physical cores. Quite incredible how games are now super CPU heavy but can't efficiently use the virtual cores.
@Frozoken10 ай бұрын
Tbf theyre not all created equally. AMD's implimentation is SIGNIFICANTLY better than Intel's in both performance and power efficiency. In fact it's essentially confirmed that Intel's upcoming arrowlake will ditch it from then on while AMD's rumoured to try 4 way hyper threading although thats less certain. Either way it makes sense as hyperthreading makes scheduling so much more complex when there's ecores and amd doesnt have to worry about that
@imnotusingmyrealname456610 ай бұрын
@@Frozoken seems like Intel is a bit of a mess
@Kyori010 ай бұрын
@@Frozoken one guy on reddit told: "The reason that SMT "works better" (i.e. gives a greater performance uplift) on Ryzen than on Intel's chips is because Intel's CPU front-end hardware is superior, and so there's less empty space in the core's execution engine for the second thread to take up."
@ronniearnold108 ай бұрын
The virtual cores are not as 'powerful' as the actual cores, though...
@Frozoken7 ай бұрын
@@ronniearnold10 True even ecores are significantly faster than a hyper thread
@shanent579311 ай бұрын
How can Remnant 2's .01 lows be higher than the average?
@ching873311 ай бұрын
Probably other way around 😅
@theh0lybible11 ай бұрын
Can run it in an average test environment with perhaps a discord call open, a few launchers loaded, and various other background apps that generally eat up CPU power. Many people utilize multiple monitors and have some kind of youtube/netflix streaming also.
@scoso13132 ай бұрын
Great suggestion. I’d love to see this as well.
@folkoven440211 ай бұрын
Is it necessary to turn off ht in the bios, or can I use process lasso and just tell the game not to use ht?
@JeremyRoche8 ай бұрын
Great video but one huge point you missed is that the CPU runs a lot cooler with HT off.
@JusH37 ай бұрын
Does hyperthreading off affect productivity tasks?
@scoso13132 ай бұрын
Good question
@voltex98853 ай бұрын
Same goes for 5800x but iam honestly unsure whats the difference bu my system with disabled smd has way less latency spikes and generaly interrupt to process latency is almost half lower so what i found out is that i can deffinitly hit more shots almost 30%more in apex legends (if i remember correctly 27.8) without smt. Fps seems same but i didnt propperly test it. Iam unsure if same goes for hypertreading but my friend has simmulare results with disabled smt. Thanks for amazing videos!
@scoso13132 ай бұрын
Thank you for all this work! Really great information and presentation.
@teskript460511 ай бұрын
nice test, can you compare only p cores with ht on + p cores only ht off against this aswell?
@jerryknudsen789825 күн бұрын
I could be wrong, but I'd bet the difference scales heavier in ddr4. All core benchmarks between the same 13th gen CPU suffer heavily on ddr4 vs ddr5 despite the fact gaming isn't wildly different. I noticed when I turned my HT off on my 13500 with ddr4 Dual rank dual channel 3400mhz cl20 gear 1 (it's cheap ram) my all core benchmarks weren't wildly different than with HT on. Infact GB6 was within 700 points on the multicore score (13000 vs 12300) That tells me hyper threads were competing with P and E cores for resources and ddr4 bandwidth was the definite bottleneck. I have 14 cores, I don't need hyper threading, but I'm about to see if I get better fps. Might be worth turning off for the ddr4 crowd more than someone with ddr5
@Frozoken10 ай бұрын
I think i noticed this same general trend on my 13600k with HT disabled in vall too. Averages onky saw marginal gains if any but in gunfights i saw my fps was dropping down a good amount less.
@RicardoMilosGachiАй бұрын
So disabled HT is the way to go?
@FrozokenАй бұрын
@@RicardoMilosGachi yes
@RicardoMilosGachiАй бұрын
@@Frozoken Interesting, thanks
@mikekangaroo112411 ай бұрын
I run msfs 2020 with a 4090 and a 10900k , I run with hyperthreading disabled as it did improve frame rate and having 10 full cores is more than enough
@lexsanderz11 ай бұрын
Sorry but the whole point of disabling E cores and HTT is to push the clocks on the P cores harder. If you running a fixed frequency you are even gimping the stock behavior. 2 cores hitting 6ghz now and then is better than all cores locked to 5ghz. Use TVB to lock push 5.9ghz on all cores.
@SoulExpension11 ай бұрын
Great vid. This is good for anybody, not just gamers. Linux users learn about io scheduler. Given the fact the OS uses a single kernel instruction queue, I'm suggesting mostly smoke and mirrors for Windows users, and the game is gonna load the gpu with work. You could make gain simply by disabling as many services as possible, freeing up the kernel stack.
@MiriadCalibrumAstar11 ай бұрын
good video but it does lack descriptive settings used in the games. Making sure that the settings are more leveled towards a CPU bottleneck(aka dont go ultra in the settings)
@viktorglibka29766 ай бұрын
It would be great to show the cpu usage in the windows task manager additionally to the fps values in each game. From my perspective as a system integrator hyperthreading increases performance, when windows deals with more threads than virtual cores present; and sometimes decreases performance, when less threads than virtual cores present, especially in virtualization and database applications.
@VnikXum8 ай бұрын
6:23 There is mistake - 1% cannot exceed AVG FPS.
@FilippoCarrux7 ай бұрын
First of all I want to thank you for your really interesting content. I've been looking around on this topic for months now and I want to share with you my experience since I don't have the possibility to afford a setup like yours. I think the 13900k is totally overkill for 1080p today so maybe you can't see so much difference because in both case the cpu is sleeping. If you have a car with 1000 hp and another one with 500hp both of the car are fine for going to the shop and buy vegetables ;) Maybe you should have tested 4k low as the 4090 could render easy lots of fps in both res but the cpu had to render double the image size at least than the 1080p or 2k. It could be interesting doing the same test with an 9900k or 10700k that are older cpu that are starting to fight in 1080p/2k competitive games that are cpu bound, with a 4090 to ensure there is no gpu bottleneck. Also, they don't have the P/E cores architeture so we can investigate if there could be a benefit also for old high core count CPUs. What I've notice in my research is that the single core speed is more important that the number of cores. If you have 12 core at 4.0ghz vs 6/8 core at 4.5/4.7, the higher clock speed do more fps in cpu bound games. Also, you don't have taken the cpu+gpu render latency that is a really handy parameter for competitive players. For example, you can play cs at 240fps, but if you have 10/15ms of latency, is a lot better playing at 180fps but with 3/4 ms. Disabling HT lower this parameter so much, so always with competitive game in mind it would have been very curious to observe how this parameter changed in this two scenarios. Finally, disabling HT you gain heat and power usage, so, with gaming and performance in mind you will have more room to overclocking. you could do this test 13900K stock -> power draw 300W 13900k HT disable -> power draw 230W (could be more) 13900 HT disable + Overclock until you reach the 300W power draw (the same as stock) With the last option, you should gain 200/300mhz on the P core and see all the benefit of disable HT. Hoping to have provided further points for reflection, thank you again for the content.
@briananeuraysem33215 ай бұрын
Would be interested to see how this might apply on an older quadcore
@teddym28088 ай бұрын
I find your choice of 1080P very strange. Sure, I guess you were thinking that makes the games more cpu limited, but that's when HT might actually MAKE sense or at least not get in the way. My experience shows that HT off gives improvement in extreme GPU load high res gaming, for example I went from 86 in RE4 remake to 95 average, and to me, that's big. 13980HX with P cores @ 5.3ghz here and a 4090m.
@Cevap0510 ай бұрын
My question here is the temperature difference with HT on/off. Anyone notice improvement with it off that’s worth keeping it off?
@mizrachayit10 ай бұрын
its cooler with HT off. My 14700K never touch 100° with unlock power limit. But if HT on, it easily hit 100°
@Cevap0510 ай бұрын
@@mizrachayit Yea I had the same observation with 12900K and HT off. And tbh with it off my gaming FPS was basically the same for most titles.
@gvido29234 ай бұрын
yes HT off is like 10C-15C less instantly im using 14900k managed to get 5600mhz and 4400mh under 70C on cinebench and around 55C while gaming
@mrcatintheboxАй бұрын
Was the core clocks the same or set manually? Without ht you consume less watts and could be boosting higher.
@rluker53447 ай бұрын
I've noticed significantly higher stable P core clocks with HT disabled on Raptor Lake. My 13900kf with a D15 and Asus Prime P is only stable at 5.5 Ghz all P core in games with HT on, but 5.8 Ghz all P core with HT off. My 13600k goes from 5.5 to 5.7 Ghz all P core stable with turning HT off. Although the 13600k also probably gets less benefit from turning HT off due to having fewer threads. It's worth checking out if you have a Raptor Lake chip.
@Ross-fo2vw5 ай бұрын
Anyone test this with World of Warcraft? I am curious because that game is notorious for 1% lows bogging down an experience when in a raid or world boss with 30-40 players casting a million spells. thanks
@morowdamien15512 ай бұрын
You could involve power consumption, since HT is using extrapower on core virtualisation
@ConorXStream3 ай бұрын
You testing starfield/ remnant/ ratchet&clank with 100% gpu usage. What difference in cpu performance you what to found
@toonnut111 ай бұрын
The video would be better with the E core's shown on and off.
@masoudfasihi434218 күн бұрын
The performance difference may be negligible but I am sure temperature and power usage wise, ht off is much better
@wolfstorm53946 күн бұрын
It's lower but in the case of a 13900k, it's a small difference that's not going to make much of a difference for you, unless you're for some reason gaming at 1080p with a CPU like this, turning off HT and the E-Cores are a waste of time except for the few games that have terrible CPU optimization that would be smoother with them off
@masoudfasihi43425 күн бұрын
@wolfstorm5394 Most of the benchmarks i saw on youtube yielded no difference in fps, but the temps were lower
@jefsonb3 ай бұрын
What about power and temperature? I tested in my system (i9 9900k) and had smaller values…
@sanchezmiki20066 ай бұрын
How do you lock you cores to stay at a set frequency i have my 13900ks at an all core of 57 and 45 but when im playing a few cores drop to 5600 for a slight scond.
@Isaiahchannel11 ай бұрын
Will this work at higher resolutions at all? 1440p? 4K? I know 4K is more GPU bound.
@saijyo85493 ай бұрын
I think most impact on 1080p, and CPU heavy load app. 4k and 1440p even lower and higher CPU just eror margin 1-5fps.
@kevinmccrea133511 ай бұрын
I personally think this is going to be the future of CPU's if they're adding all these extra cores..
@impuls6011 ай бұрын
Heat consentration looks to be a huge problem as we go smaller. The new rentable cores leak from Intel seems to support this.
@shanent579311 ай бұрын
The future is fewer cores and more threads. Intel is just going to change the names and pretend that a thread is a core
@shawnarwood11 ай бұрын
@@shanent5793you mean bulldozer
@shanent579311 ай бұрын
@@shawnarwood kinda but hyperthreading can use all of the units on a single thread. Bulldozer had two integer cores that couldn't both be processing the same thread
@jakedill130411 ай бұрын
Actually the future is.. different.. Intel is reworking how they do hyperthreading as a concept and basically getting rid of it in the form that you know of and we know of now.. it's actually kind of the most exciting thing that Intel's done in a while, but won't probably see the light of day for another year or two.. maybe even longer for consumers.. but it is cool and makes a lot of sense.
@Remortality-6 ай бұрын
My cooler is not the best so I can run my 13600k with hypertreading on 5.5 pcore 4.3 ecore and 5ghz ring ratio without thermal throttle but when I disable hyper threading I can run 5.8 pcore 4.4 ecore and 5ghz ring ratio without thermal throttle not sure what I should run since I mainly use my pc for gaming
@gvido29234 ай бұрын
less temperature and higher frequency is way to go and 15th gen cpus wont have HT anymore
@mannyc1911 ай бұрын
This holds true for Intel, not so much for AMD Chips like my 3600x and 7800X3D systems/
@mattzun67792 ай бұрын
Testing with and without hyperthreading on a 16 core CPU is somewhat pointless. We KNOW 16 cores without hyper threading is enough for gaming because the 8/16 core 7800x3d is the gaming champ. What would be interesting is a test of the 14600k With 6P cores and 8 ecores, it would be a great test of whether hyper threading is actually useless for gaming.
@Lunnyx5 ай бұрын
For the new intel gen, the best thing to do, is to disable E-core when gaming, by doing this i improve my average fps and 1% low by 15%.
@Darksh1n8 ай бұрын
You dont even need to go into the Bios to turn off Hyperthreading you can do it in a program called Process Lasso depending on what game you play.
@H0tCh0c0-gr3qe8 ай бұрын
nah
@MrAyybee2cold8 ай бұрын
You do it thru bios because process lasso will be detected by easy cheat in multiplayer games.
@MrChefft7 ай бұрын
For CS2 use -threads 9 on launch options with that cpu im using -threads 7 with a R5 7600 and have nearly 30% gains on low's and average, insane
@vladislavkaras4918 ай бұрын
I wonder how performance can be different, if we have some tasks in background, like Discord, browser activity, downloading something and so on. Would the performance difference be bigger than what we saw in the pure game benchmark? Also, would be cool to see the same test on AMD platform! Thanks for the video!
@Frozoken7 ай бұрын
Doubt it seeing u have ecores to handle that. Thats part of it ecores made HT useless and HT also makes scheduling thereads when theres also ecores wayyyy more complex
@teddym28088 ай бұрын
Also, disable core parking with the free park control app. Takes 2 seconds.
@michaelthompson979811 ай бұрын
I can only assume that having HT off (in some titles, and certain game engines) getting better fps as a result of power budgets being redirected into better performance of the P-cores and giving them better higher boosting clocks = higher fps.
@Frozoken10 ай бұрын
Could also be from songle ecores from each cluster taking some load that hyperthreading would have because not only is an ecore much more performant than a hyperthread but theres also only a core to core latency penalty between ecores in the same cluster. The cluster of 4 all have to also share l2 bandwidth so ipc is significantly higher when loading up different clusters first and the scheduler does do this thankfully. I also tested this in a gamewith process lasso and found that two ecores of different clusters were significantly more smooth than one pcore at 5.8ghz with the ecores being at 4.5ghz. Usually its meant to be like 2.5-3ish ecores = to one p core especially in something like faming but just 2 were still noticeable ahead when using different clusters and as i saod, much less core to core latency penalty like that too
@FireJamUSA5 ай бұрын
I'm running a 13700k / 4090 and turning off HT made Hogwarts run smoother. Thanks for this video :)
@wjjohns04able9 ай бұрын
I’m not sure it all just about fps. Some games I play run a lot smoother. Maybe it’s got more to do with game integration and how it’s coded? 5fps one way or another doesn’t matter much to me, but in fps if it’s not smooth that’s a big problem.
@khaledalhamdan26279 ай бұрын
thank you for really the good video, and it is good that you included World of Tanks in the test as it is CPU heavy game.
@wawaweewa915911 ай бұрын
Noice, i wonder what getting rid of hyper threading, now that we have plenty of cores available in cpus will do to die size and complexity of design
@zodwraith574510 ай бұрын
I didn't even need to watch the video to know the outcome. Hyperthreading is one of the most misunderstood technologies in the hardware space today. While it was very beneficial when we were on 1-4 cores it's given diminishing returns since then, and is pretty much obsolete today where 8 cores are standard. Especially when E cores perform many of the tasks that made HT most beneficial in the first place. It's no wonder Intel is dropping it in future generations. In today's CPUs where the biggest hurdle is fighting the laws of physics any way you can shorten that pipeline should theoretically be beneficial. This has me interested in what Arrow Lake will bring.
@astralcomet110 ай бұрын
I have the exact same ram kit and a new 14900k. Can you tell me the specific Ram timings you are using? I plan on taking the heat spreaders off and putting a fan over them to cool them 👍🏽 Mobo is the msi z790 pro a btw
@astralcomet110 ай бұрын
I also plan on disabling the e cores as I mainly play valorant and cs and want the max latency reduction that’s stable. For Ram stock Xmp I’m at ab 56ns aiming for ~35ns
@Frozoken7 ай бұрын
@@astralcomet1 Dont disable ecores for valorant ive tested it, they slightly help and for CS they actually help quite a lot. If u really want low latency with the ecores disable all but 1 from each cluster of 4. Ecores only have a latency penalty when talking to each other in the same cluster and not only that, with only 1 activated per cluster ur getting near 3dvcache levels of cache per core so theyll almost definitley be significantly more performant for gaming.
@StroggKingu11 ай бұрын
I wonder how it would behave in Cyberpunk as recent patch added SMT option to set.
@cblack347023 күн бұрын
You do get more fps with it off but you will run into stuttering and screen tearing it's better to have it on and have a constant stable frame rate, not worth the extra 3-5 fps
@FreshJ1v311 ай бұрын
HT of or off for New World? Any help please and ty. Azoth Engine is strange!!
@leandrodrace11 ай бұрын
if u have 8 cores or more yes.
@doctorsatan426011 ай бұрын
Currently only 3 minutes in video: hope this will answer... On a traditional cpu like i5 12400 is 6c12t always better with it left on, especially with browsers, medal capture, steam, game, discord all open while gaming with in game voip, and discord running.
@takchuewong51446 ай бұрын
it's not about performance or speed. All games just crash when HT is enabled on 13900k. Devs aren't going to study on this old school tech and just told us to disable the e-cores.
@AshtonCoolman11 ай бұрын
Do the same with AMD's 7950X or X3D
@SkyBeyondSkyVM11 ай бұрын
on 13:00 what game you playing?
@DannyzReviews11 ай бұрын
Baldurs gate 3
@SkyBeyondSkyVM11 ай бұрын
i see Thank You
@IPowder41059 ай бұрын
I think you should try on 13100 to se the difference because that's when cpu will be used a lot more
@TurboD16z621 күн бұрын
The dumb is, the most important things are power and temperature difference. But again all is so focused on fps only.
@Randalsgaming9 ай бұрын
Hyperthreading just adds to pc latency , I was running apex at 210-240 fps with it on with a i7-11700k and a 3070, after turning hyperthreading off I get a solid 240. I know the i7-11700k is a bottle neck for the 3070.
@tomtomkowski76537 ай бұрын
What about the power consumption and temperatures? It should be much better.
@Frozoken7 ай бұрын
I have an overclocked 13600k and under full load it went from about 240w to 200w with hyperthreading off. Gaming was like 10w difference tho.
@PAIN-ot4cj9 ай бұрын
Danny can you so one like his on the 14900k or 13900k with rebar on vs off
@LorentGuimaraes4 ай бұрын
This is missing average temps.... sadly
@echickentiken92007 ай бұрын
How mini core off?
@auritro39037 күн бұрын
6:16 1% lows is higher than average FPS 💀
@nivea8788 ай бұрын
i encountered much less stutters whendisabling HT (14700K)
@isaaccoido6 ай бұрын
can you try this with helldivers 2 ?
@kirkh42059 ай бұрын
Nice report, I would be interested in your temps during these tests too. Does having HT off help lower temps at all?
@v0rtex878 ай бұрын
ht off significantly descreases temps, which author doesn't said about
@Sn4p_Fps5 ай бұрын
Temps are the same in some games less in some games higher , nothing to worry about but the % whit ht off is a lot higher
@LuXoR15906 ай бұрын
Cpu is cooler when ht is off
@nasunabi6 ай бұрын
i have i9 10 something something , i turned off HT and instantly gained 30 fps in COD warzone
@oskarteran336511 ай бұрын
but in core i3 12100?
@Shuttterbugg11 ай бұрын
Good video! Alot of pwoole don't know to get.max wprformabce u can set up your bios to turn off ecores by click of a button like scrlck
@AleksCoreBY7 ай бұрын
But disabling E-cores doesn't improve performance. There's actually no reason to disable it unless you have very bad cooling
@TheGamingSKITZ11 ай бұрын
Not the video i was hoping for. If i wanted to look at grahs id go to a site. I was hoping for a side by side comparison like alot of people do
@jessefisher180911 ай бұрын
I thought I heard intel is getting rid of hyperthreading going forward, something about it getting in the way of future development.
@shanent579311 ай бұрын
They're gonna call threads cores and pretend that HT doesn't exist, just like it's 1984
@Frozoken7 ай бұрын
@@shanent5793 No it's literally just going to be disabled lmao
@deaddeyedante11 ай бұрын
Great video.
@Seu_Lunga10 ай бұрын
Do you have this amount of games tested for v3 or v4 Xeon popular x99 platform?
@blindman80936 ай бұрын
csn you try it on warzone 3
@DualPerformance11 ай бұрын
Yeah sure, but try with an I7 2600K and you will only get more FPS with HT On in 99% of the games released in 2022 - 2023
@Waldherz11 ай бұрын
This video is about MODERN hybrid intel CPUs. Not 10+ year old CPUs.
@jakedill130411 ай бұрын
Right but you're more likely to see benefits from turning hyperthreading off on older hardware if it can manage the basics... Applying this only to modern current or just well as of a day ago recent generation high-end Intel consumer CPUs is kind of a waste in general as that's such a minor use case.. and as far as hybrid CPUs go.. the premise wasn't it's hybrid nature that was the problem so much as it was whether or not the hyperthreading mattered in the performance cores... Not that it matters, given this person was pointing out that there's a point in the architecture history where the scale goes the other way... Which is supplemental, not contradicting.@@Waldherz
@Waldherz11 ай бұрын
Well, about the Hybrid architecture. People started asking if you could disable HT (in previous videos on this channel) since the total thread count is so high to begin with. I personally dont get why I would buy a modern high thread count CPU to then disable some threads. Just get a lower tier CPU instead.@@jakedill1304
@MikkoHiiri11 ай бұрын
Intel says that 14th series is up to 16% faster in metro exodus than 13rh due to optimization. Does 14th run it automatically without E-cores? 🤭😅
@dighekonesijbeg549810 ай бұрын
Age of Empires IV? Pretty please?
@yoked39110 ай бұрын
Cs2 250 - 300% difference from 1% lows to average fps is nuts 😂😂 dumpster fire, i can see why it feels like crap
@riki-s3 ай бұрын
Far cry 1-5, dota.... loves ht off
@jeffreybouman211011 ай бұрын
I had a 9700K Switched it for a 9900K Lesson: no performance difference at all Only in 3DMARK tests the different was big but in real gamjng it was the same or the 1% lows where a bit worse
@quademasters24911 ай бұрын
These days I only look at single core performance. If single core performance isn't significantly higher, you probably can't feel any difference.
@jeffreybouman211011 ай бұрын
@@quademasters249 amen 🙏🏻
@Cosmicstring-dg5iz19 күн бұрын
yeah,but threads on 9900k are twice than on 9700k, and it's matter.
@dn77836 ай бұрын
Can you test it with fixed clocking
@MsTatakai11 ай бұрын
I tested on my 4770k and NO... way slower ehehe XD
@club4ghz6 ай бұрын
You need at least 12th gen intel 6 cores or a high core count older cpu to disable HT in 2024
@jakedill130411 ай бұрын
LOL older game engines that rely on single threaded performance.. like unreal engine 5.. oh wait excuse me the script reads unlike unreal engine 5.. can someone tell the CPU that please? No? Well.. oh well I guess...
@YourPalHDee10 ай бұрын
Well yeah it's pointless on a 13900k 😂 its totally valid on any i5 equivalent. What a silly video.