God thank you so much for your content. You're straight and to the point, extremely informative, and just so professional. Glad I can watch videos about this stuff and actually LEARN something without having to sit through dubstep intros or someone shilling a sponsor for 2 minutes beforehand.
games are very cache sensitive. these two cpus that were tested had very different amounts of cache. the information in this video is really describing how much L3 cache matters for gaming. too bad we didn't learn what we thought we were going to learn.
Finally glad to see someone do some real testing on this. After seeing how badly 4C/4T i5's fell flat on their face compared to their quad core i7 counterparts, I won't buy a CPU without hyperthreading anymore.
@@TechDeals That's the exact reason I chose to upgrade to the 3700X over the 9700k on Black Friday. When the 9700k wins, the Ryzen chip still hits high enough and consistent enough FPS to feel smooth and responsive. When a game wants more threads than your CPU has, it's a stuttery mess and feels unplayable in many cases. As an aside, I also picked up a 2TB Intel 660p SSD to go with my new CPU/MOBO, pretty much entirely based on your testing. Thanks for all the great info you give us :)
@@DrearierSpider1 The 660p really is the best value option, there are plenty of other drives that are quicker but they cant hit that price/gig ratio of the Intel.
@@dave-ish8098 Ya, that's my justification. My rig is really gaming only, and gaming isn't very storage intensive at all, as long as they're installed on any basic SSD. There are use cases where going to a 970 Evo or 970 Pro make sense, but those are more uncommon than common. Considering games are hitting even over 150GB now, capacity is a much more important consideration than raw performance for that use case.
Battlefront 2! A lot of people still play it, it's gorgeous, fun & taxing on low-end & midrange systems. It does DX12 & DX11. Please!!! Great video! I was excited to see this topic done again. Especially with your testing, which I always respect.
Man I'm so unhappy with my 6600k and wanted to move up to a 7700k but the problem is price its cheaper for me to go build a new ryzen for a longer-term then add that processor. Good video tech deals
@@nichohells Yeah same position and I'm delidded @5ghz so I guess the smart play is to wait till the next releases. Might even need to wait until 2021. Oh well, first world problems :)
If you can get lucky and find a X99 Motherboard on a good deal I can't begin to tell you how much fun I'm having trying out all of the different CPU choices their are for that platform when you include the NOW INSANELY CHEAP XEON'S... They are even good prices on them on ebay but finding a 8core 16thread Intel CPU for under $120 ... soooo good and even 10core 20thread models for under $150 as well as the 12core and 14core but now you are getting up into CPU's that have too low of a clock to be good for gaming really. If anyone really loves to tinker you should check out that platform or if you are looking for something even more budget then go for X79... Quad channel DDR3 clocked high really has no tangible benefit in gaming over DDR4 in dual channel ... and there is even an argument out there that it could be better as far as latency... even the way the X79 CPU's architecture is built its possible that you will see lower input lag from your mouse and keyboard if plugged into the right USB ports on X79! It's definitely something to look into.
Hunt Showdown is a great example of a game that has dramatically better 1% lows with a higher core/thread count. The community benchmarks for 5800x versus 5950x show a huge uplift in the 1% low figure, I was shocked. Usually most benchmarks don’t seem to show a big difference, but in my experience at least some games do benefit in terms of “smoothness” with more cache and more cores and the difference is night and day in such cases.
Hi Tech, huge fan here, love the way you scrutinize every detail of whatever subject you analize and give us an informed and concise conclusion at the end. You asked for sugestions about games to test and I belive that for a CPU test it would be interesting to test Red Dead Redemption 2, it is quite demanding and is probably the closest we have to a "next-gen" game (not really but you get my point). Also Control is probably a good GPU benchmark. Apart from that recent demanding games are the best bet. Well thank you for your time and hope you have a nice day :)
Awesome video! The editing is very well done, and lots of good information. I've used so much of your content to learn about PCs and when I am ready to build my first PC someday I can say you truly played a big part in it! Keep up the great work!
I'm amazed of the quality that this channel offers right now. I remember watching videos from 2016 or so, 1070 SLI vs 1080, etc and the quality of the videos has improved massively. Thanks for all this nice videos! Regards from Uruguay! :D
I absolutely love the new editing as a result of the new editor. Makes the videos more fun to watch and even more professional! Keep up the great work!
You understand everything. Thank you for not using a 2080ti. I do a fair amount of testing myself and have some games i could recommend. Cities Skylines: With a population of 100k. More cores tend to scale well with simulators in general. Crysis: Classic and it also scales with more cores. Despite its age i still use it for testing hardware. Deus Ex: Mankind Divided: Has options for DX11 and DX12 which both behave differently. GTA V: Popularity and likely a legacy benchmark like Crysis. Battlefield One or 5: With 64 player maps tends to be pretty aggressive on the CPU. I still can't get Watch Dogs 2 to run properly. Likely because its a port. RDR2 might be an interesting one too. Just some thoughts. Keep up the good work.
Regarding the AC Odyssey results, I've seen this before. I saw a benchmark where a 4 core Xeon with HT was being compared to itself with HT off. I AC Odyssey, same phenomenon. The thing was that the CPU was so busy it just wasn't loading things in the benchmark and wasn't switching to more detailed models when it got closer to them. You can see it with the LOD taking a while to change the texture on the roof there with the 1300x, in the begging of the benchmark. Therefore, better performance. I've also seen this on the AC Origins benchmark with a Pentium g4560, where it just gave up and the physics broke; no animations, things going through each other, NPCs not reacting to things, texture quality not being updated, and way higher framerates, because the CPU wasn't calculating anything.
I could be wrong (entirely mostly possible), but one advantage of having HT is when you're streaming/recording games. The extra threads allow other things to be going on while you're gaming. Except for games like Shadow of the Tomb Raider. My 1500x was okay with averages of 50-60 fps, but had turn a few settings down. It wasn't my gpu. The CPU was maxed out, while the GPU still had plenty of power. Changed out to a R7 2700 I could get about 10-20 fps higher.
I prefer Tech Deals to Gamers Nexus, Linus Tech Tips and Hardware Unboxed, though I love them all. The clarity and humanised approach is far more palatable for a tech noob like me. Clear and concise. Love the new editing; and haircut.
It is very dishonest to call new dawn DLC. It is a vastly different game to 5, even the map is really much different. People are just morons and like to complain.
18:40 - Empirical evidence of HOW much SMT helps with few cores CPUs 4/8 running like a 6/6 if not better SMT is almost a "must" for people who want to play and making other things without bothering anything
it's also empirical evidence of how much L3 cache helps with games. i wish someone could do a test where the cache was the same, and just the SMT was the only difference being tested.
TechYesCity did a video on the 3500x - a 3600 with disabled SMT that you can only get on AliExpress. It's really interesting that the 3500x actually did better than the 3600 in quite a few games. That being said, if you compare the 7600k which is already on the way out to becoming unusable and you compare it to the 7700K which is still holding strong, I wouldn't be surprised to see a similar picture in a year or two after the next gen consoles launch with 8 x 3.5 Ghz Zen 2 cores.
For me it's not a matter of fps difference. It's multitasking. Playing a game, having discord open, recording gameplay, having chrome open in the background, etc. My buddy has a 4c/4t intel and can't even talk in discord while playing some games because his cpu is pegged at 99% the entire game. It matters for much more than fps imo
Your channel is amazing and i would say it provides a lot of scenarios for an average and low budget gamers about their builds and expectations . Keep pumping up these contents they are amazing and very informative . :)
Tech Deals thank you so much Mr, Tech, i hope you see this, you can use the Replay feature on Fortnite and Overwatch to get the same game for different hardware, i love you videos 👍
A even simpler way to explain Hyperthreading / Simultaneous Multi-threading (SMT) is hardware implementation that makes your 4 core processor looks like that it has 8 cores, without actually implementing 8 physical cores. So operating system / software thinks it has 8 cores available to do tasks. In terms of performance scaling, it is not quite double the (8 physical core) performance. But it'll do better than having just 4 cores. (The above assuming 1 thread per core)
Well things that work about this edit; how the large graphic on the bottom of the screen came in and left the screen at 1:15. How the " amazing value" graphic came on to the screen at 3:45 also the "can be upgraded" soon after these are animations that are better for presenting information to a viewer. The flashy animation used for showing processor names a like in the first 3 min should probably switch roles with the one used for the "amazing value" graphic, as that flashy animation should be reserved for key points as it can get old fast. I am probably being picky but I figure you just hired an editor, so I wanted to get my two sense out there early, because course corrections are less severe earlier than later. I overall just felt like it was over animated in general with flashy animations for set up information and more basic animations for key points. It's a strong case of situational irony from my point of view that the key points of discussion only stood out from other points in the conversation because those were the points that received the least enthusiastic animations, even though it was verbally the most enthusiastic part. Videos, even informational ones should have a natural flow to them and these animations were to make a science reference an example of destructive interference where you clearly had one flow to your video, and the animations tried to come in and fill in the troughs and level off the crests of your flow. I do really enjoy the content and its exciting to see you moving on to this next step and now that it is out sourced the more you need to develop editing standards mostly for channel continuity sake. If you look at linus his videos feel like they were all edited by the same person this was a sharp change from the norm, yes production quality technically went up but the animations almost dulled the story more than they helped you tell it at least before the benchmarks. Hopefully with time things will iron out feel more continuous.
Been subscribed to Tech Deals for a while already but missed a lot of videos. Upon checking this video out, man he has really gone a long way! The way the video transitions to other scenes and that intro alone says a lot! Love your videos, Mr. Deals! Much love from the Philippines!
Honestly, you had a lot of cock-ups in this video. But it's a really valuable vid because you decided to go ahead anyway and it really demonstrated the complexities that are often hidden. It wouldn't have been half the vid if you re-did it to perfection. Top stuff!
Hi tech! i had the same problem that youre having with the HT. If you set the monitorization of msi afterburner to each thread instead of the overall cpu usage you will see that Windows sometimes dont use the actual physical cores. If you manually set the afinity to the physical cores the gameplay will be smoother. I reallized this when playing old games that dont use many threads and when streaming. I have a Ryzen 5 2600. Let me know if it works for you. Love your videos man, greetings from the other half of the world c:
that is when 4 cores aren't enough but what if its a 6 or 8 core cpu does hyperthreading cause instability or worse frametime or doesn't make any difference
This video has arrived at just the right time. I'm thinking of upgrading from a R5 1500x. It's all food for thought as I do notice some stumbles in Shadow of the tomb raider.
One of the things that immediately jumped out to me in the benchmarks (specifically for The Division 2) was that while the fps numbers were similar-ish (with the 1500X in the lead, naturally) the 1300X system was completely CPU bottlenecked. Look at the CPU/GPU usage numbers during those tests and you'll see the GPU was only around 40% usage while the CPU was pegged. The 1500X was closer to 60% or so CPU usage, but the GPU usage was in the 80s...
Hey, I like the new intro! Almost wish it was just slightly longer. But I can understand the need for brevity when the rest of the video has to encompass Tech's take.
Hey Tech, I don't play Steep, but I would like to recommend adding in Star Wars Battlefront 2 to the benchmarks. Or Battlefield V since they run similarly. I know you throw them in sometimes, I just offer that it's makes for a good CPU test. I have a Ryzen 3 1200 system in the house and a Xeon X5675 system. The Ryzen system runs the game 75-110ish fps with a fair bit of stuttering, while the Xeon runs it 80-150fps (and occasionally as high as 180-190) with almost no stutters at all... Pretty smooth for the most part. These were both using a GTX 1660 TI and 1080p medium-ish settings. Admittedly, there could also be some memory scaling going on, as the Ryzen has 12gb in single channel, while the Xeon has 24gb in triple channel. Newer isn't ALWAYS better. Also don't ask me why I have this random assortment of parts. Lol 🌟 🌟
Dig this guys videos. He covers all the bases he can think of, explains things in detail without going too far (Like mentioning the recording of the data did not affect either cpu as it was sorted externally) and is maybe less "Fun" than watching Linus drop everything he touches or Jays 2 cents comedy, but ALWAYS an absolute professional! I also see no FANBOY loyalty either way, just the facts jack! Great videos!
Buying a New Processor today makes you feel hyperthreading isn't important ...... Tomorrow is when you wish you're Cpu had hyperthreading = when your Processor starts to struggle with newer more demanding games. When i upgraded from i5-2400 to i7-2600 after three years, it was night & day ! Kinda like going from 8 to 16gb of system ram or from a hdd to ssd 🎮 either way, Thanx 4 Posting 🎥
Hi, I'm using an i5 6600k and recently I've been thinking of changing it for a 6700k or a 7700k. Do you think this is a good idea? I just have to change the cpu. The rest of the pc is compatible. I've seen some i7's for arround 200€ and I think this is quite a lot for this CPU's. Thanks
There is a video on the channel doing just that! We upgraded an i5-6402p to a i7-7700K and replaced the cooler. Nice upgrade, but Ryzen is a better deal, even with the motherboard change, unless you find a cheap CPU
The editing just went top notch with a minimalistic approach. While I feel like it's work in progress, next I would really love to see a better table graphs. They really look basic and boring. Just a constructive criticism. Keep it up!
Just want to say I really liked your new vibe!! You are as Knowledgeable as the top tech youtubers. I respect you so much, and really hope this really pushes your enthusiasm!!
Great video as always. There is a bit of echo though . I would also advise something for the background other than just white, for a more visual appealing scenario
Thank you for including TW3K in your benchmark, i play more strategy games so i really appreciate it. :) Hope you'll include at least 1 benchmark in the future for the strategy fans.
I can defend the CPU choice here. Right now, quad cores have the most exaggerated difference in game smoothness depending on whether or not SMT is turned on. In a lot of cases, it can take a game from a stuttering mess, to nice and playable.
@@Cakerpie That's silly. If you compare something, you want to eliminate as many variables as possible. This video should be titled "Ryzen 5 1500x vs Ryzen 3 1300" since he's comparing different products, and not just how SMT affects performance.
A suggestion: label the graphs differently for live gameplay and in-game benchmark. Live gameplay results have more uncertainties -- as you pointed out in the video several times. But viewers jumping between graphs may not be alerted as much to the voice over explanations.
I would recommend you check out Vermintide 2 for CPU/GPU testing. This game is heavily CPU bound, better yet I noticed clear differences between i5 3570k and i7 2600k with HT. In that game there's an AI that calculates the position of players and such, and when there are hordes of enemies about to spawn, it usually hitches/stutters on lower end CPU's. Another interesting HT test would be turn time in Civilization V/VI as it's known to be a good test.
Hyperthreading rocks Chrome, Sony Vegas, and other apps like Adobe. Problem is Intel has ancient insecure shit in their CPUs compared to AMD. Chrome and general computer usage benefits from treading but Intel is insecure as shit and will keep showing down when updates from Windows are implied
It would be awesome to see benchmarks for games like "Arma 3", "Squad", "Post Scriptum", "Green Hell" and "The Isle". Please Tech and Rogue? :)All these gold stars Ive earned gotta count for somthing lol. If you need help in any way with the games I listed contact me. I can get you in a game with my group, we can teach you and do anything you need to get the benchmarks. It would take planning but we could have up to 80 people participating in creating benchmark scenarios on Post Scriptum. Please check them out, independent developers have made some amazing games in recent years. AAA titles make up a small portion of my library. Thanks again for all work you do! P.S. Rogue you might particularly like "The Isle".
Your videos are awesome!! I like them, you're doing a great job! I would like to suggest Total War series games for benchmarking, they are veeeery demanding games, even the old ones (Rome total war, medievla 2 TW), my favourite is Attila Total war, so I'll suggest it. Greetings from Peru!
Rogue: "You've been playing video games all day! Get off and hang out with the family!" Tech: "No I'm not, I'm making content for our RU-vid channel!", with a smirk on his face.
Lots of people still play GTA V. Matter of fact if you look at steam stats you'd be surprised for a game that's 5 years old. Would love to see it included in future benchmarks.
SIMD/Hyperthreading can't fix everything. But... at times it is very useful. And a smart way to get more performance in MANY situations. So yeah. It isn't a perfection solution...and good to see this video so people can see WHY you can't just expect double performance. Or HOW you get the performance. Now the moment you actually need to compress/edit a video for youtube though... those extra threads really begin to speak out.
I've been with Tech Deals since 2016. I've always used his affiliate links and haven't been led astray. Using his affiliate link for the first time from 2016, I purchased a Samsung 960 gb SSD hard drive back then for $200.00. Few dollar savings here and there adds up real quickly......like a footlong Subway sandwich w/chips and drinks "quickly".
I've been on the fence for a while about the RX 5700 XT vs. 2060 Super for my new rig. I game at 1080p75. The info in the beginning of this video about the shortcomings of the AMD card at 1080p was very helpful to me. I know that wasn't the reasoning behind this video, but thanks for the info anyway.
HEY there I really enjoyed this video as I'm about to do my 1st build... Going from a 2013 I5 3570k 4 core 4 thread to a Ryzen 2600x 6 core 12 thread while keeping my current RX580 Gpu. Then reassess my Gpu requirements after that. I've been playing Sottr and GR Wildlands and have bn putting up with glitching even on med/high settings even crashes to desktop as the cpu hits 100% for long periods. I've chosen all my parts based on your videos and really appreciate your advise. Many thanks mate from New Zealand.
Hyperthreading rocks if you love technology and want to code, run virtualization or be a 3d artist. I loooove technology since I was a kid in the 1990s and hyperthreading rocks for running visual studio code from microsoft.com and doing Hyper-v virtualization on Windows 10 pro. I have 13 Virtual machines with 2 virtual routers (pfsense go Google) and 11 virtual machines with Windows Server with Hyper-v on Windows 10 pro. Hyperthreading makes this possible. Google it if your interested in technology?
Feels like up to a certain number of cores hyper-threading often (not always) helps. If you have 8+ actual cores or more there is a rapidly diminishing benefit - at the moment.
Well many many comments on this video thanks tech you are going to make me type an essay in pieces throughout the day. First off I enjoyed the subject matter. Second I understand why you would use the 5700xt, on a channel called techdeals. I feel it's worth showing that the card and the quirks that come with getting the best deal you can. Third no real opinion on steep, I think it's fair to include some games that normally aren't benchmarked other places for the sake of variety, but I also don't know how much it would be played by those watching your channel. So umm straw poll on a few live streams maybe?
Not for most games..if any. Right now games care about cores and core speed. However many programs speed up when using hyper threading - e.g MATLAB, parallel computing package for Python or R.
The real telltale is in showing the % drop from average to the 1 and .1 percents. That is the best way to tell the smoothness in a game. I am playing most of these games on my 2600K @ 4.5ghz and GTX980 with decent performance. Some games I need to lower the settings which are more cpu intensive, but none of these games are unplayable for me... That said an upgrade would show a big improvement in some titles for sure. One of these paychecks that will happen.
A bit unrelated. But I wanted to ask for your opinion. I am currently buying parts for my first ever PC, I prefer getting parts at my local best buy rather than getting stuff online. Is buying the Ryzen 5 3600 worth the extra $80dlls at best buy rather than getting a Ryzen 5 2600 on amazon for $110?
