add arrow-down arrow-left arrow-right arrow-up authorcheckmark clipboard combo comment delete discord dots drag-handle dropdown-arrow errorfacebook history inbox instagram issuelink lock markup-bbcode markup-html markup-pcpp markup-cyclingbuilder markup-plain-text markup-reddit menu pin radio-button save search settings share star-empty star-full star-half switch successtag twitch twitter user warningwattage weight youtube

Comments (Continued)

  • 6 months ago
  • 1 point

Well in my preceding post I didn't make any claims about games, I'm just arguing overall computing power.

Most games use 2 cores maximum

This going to be a problematic assumption for several reasons. First I don't believe you have any data to back up that claim, it's just a repeated idea originating from the early days of multi-core systems and people wondering over the value of getting a quadcore (or tri-core, remember those AMD cpus?) over a dual core system. Secondly, multi-core programming has matured quite a bit. True you can't make any arbitrary code or program necessarily run an an infinite number of cores, at least there's not always value in doing that. I'm not sure what you know about multi-core programming, but in generally the developers aren't manually deciding which core a bit of code is running on and only using two cores. Games may not strictly require more than two cores, but if the work they're doing can be split across several cores they will utilize that, if available, as much as they can. It's one of the nice things about multi-core programming, it is flexible in that you don't need a fixed amount of cores to do a task. There's plenty of software that will run just fine on a dual core system, but will be able to utilize a Threadripper high core count just as well. And at the end of the day games are just software, not special case exceptions. Again they don't necessarily need or benefit from (super high) high thread counts, but there isn't a magic value you can arbitrarily assign to games or "most" games either. Clearly we want more than two cores for gaming now, and at a certain point, for most users there's dramatic diminishing returns from adding more cores and more threads beyond a certain point. But at the very least it's better to have more cores/threads than you need than to have to few.

And I'm just of the opinion that paying the premium for a 9900k and disabling hyper-threading so you can get a couple hundred mhz higher OC because HT "isn't needed" isn't a practical solution, it's an enthusiast solution. Overall and in the long run you'll be better off with the additional power HT provides over a couple of hundred mhz in OC. At least I would be.

  • 6 months ago
  • 1 point

Right but enabling / disabling HT is only a few clicks away in bios, so what do you have to lose?

If you play, and only play, Arma 3 or Dayz, for example, which many people do, you would be better off with HT disabled and a faster OC. This can specifically help not only FPS but the frame latency.

"but there isn't a magic value you can arbitrarily assign to games or "most" games either. Clearly we want more than two cores for gaming now, "

You speak as if the 9900k is a dual core cpu with HT disabled. Did you forget about the other 6 cores?

BF1 is THE GAME that supposedly, the more cores or threads the better, but the 9900k actually performs worse than the 9700k, despite having 8 more threads and more cache. Not only that but the frame latency is almost half; 4.7ms vs 9ms in some cases. https://www.youtube.com/watch?v=568uEmI-et4

GTA 5, 9700k 108fps 9900k 115fps, however 9700k always 1 second of latency less at the same fps. https://youtu.be/568uEmI-et4?t=85

The difficult thing to know is whether the extremely slight FPS increases in favor of the 9900k are due to the extra threads or the extra cache. My guess would be the cache because HT introduces latency which you can see in the video on each game.

4.7ms vs 9ms is a BIG difference on a 1ms gaming display.

  • 6 months ago
  • 1 point

As far as your comment on "Most games use 2 cores maximum" kinda falls apart. Years ago with games like GTA V and other triple A game titles from that era has been proven to fully use 4 CPU cores. They were often proven by showing FPS differences between the i3/i5/i7 of the time and it could peg a 4 core unlocked overclocked i5 to 100% on all cores. These days games like Assassin's Creed Odyssey on benchmark videos where it can slam a 6 core i5-9400 at 100% use all cores. Can't say that is all background tasks but games now are using more and more cores as game developers know people are likely to have them.

Before 2010 the logic was that most games could not use more than 2 cores but times have changed.

Now the differences from the current 6c i5 and 8 core i9 are not huge and an argument can be made about cost to performance whether the i9 is even worth it as you can still have a great gaming experience on a mid range 6 core CPU from either Intel or AMD.

As for disabling HT for added performance it really depends on the tasks that you are doing. Your examples do seem rather far off too. 111 FPS is on average 1 frame every 9ms and 212 FPS is a frame every about 4.7 MS. Sure displays having 1MS response times are nice to reduce ghosting but saying disabling HT does not double your FPS. If your PC is pumping out 144fps then it is delivering a frame about every 6.9ms and that 1ms response time on the display allows most of that frame to be fully visible with limited ghosting from the last frame. Well it is true you might be able to see the difference of 111fps compared to 144+fps on a 144hz display with a 1ms response time but not on a 60hz display with the same response time.

  • 6 months ago
  • 1 point

Well besides 1ms response times on monitors are nonsense. It's a specific number cherry picked by the manufacturer and in no way represents what you're going to see in any real world use. It is usually the fastest GTG time they can dredge up, and not even the average GTG and no where near the slowest. So the idea of doing real world performance math with a 100% pure baloney marketing number just doesn't square with me.

Maybe I'm just old, and remember when LCD monitors were still actually terrible for gaming, and people were going to give up their CRTs when they were pried from their cold dead hands. But seeing how this subject has evolved has been fascinating as response times have become this thing people twist themselves into knots over. Having been a user through it all, I have to say it does seem like people can't be happy and if they don't have a problem they'll make one up.

Or maybe I'm just fortunate that I can use a decent IPS panel without torturing myself over a few ms... thank goodness.

  • 6 months ago
  • 1 point

I went from a 5ms panel to a 1ms panel and yes I am aware that the MS rating is cherry picked. Though the cherry picked 5ms is definitely slower than the cherry picked 1ms. This is a good website to see ghosting and it makes the difference between the screens easier to see. I do see the difference in fast paced games as well so it does have some value. I do agree though to take it with a grain of salt.

  • 6 months ago
  • 1 point

I don't need a website to see ghosting. Maybe you missed my anecdote

Maybe I'm just old, and remember when LCD monitors were still actually terrible for gaming, and people were going to give up their CRTs when they were pried from their cold dead hands.

I've seen legitimate ghosting on early LCD's. I lived through a time when they weren't really viable and gradually became more so and eventually the defacto standard. For me the technology become good enough over 10 years ago. Improvements are nice, which is why I prefer IPS screens and adaptive sync technologies, and fortunately I can live with the "slower", albeit extremely bearable, response times for better picture quality and viewing angles.

Nothing is quite perfect, just depends on where your values lay when choosing your compromises. Fortunately the compromises are getting smaller and smaller too. But that's another reason when someone comes in claiming BIG differences over a couple of ms I have to chuckle. Because, hold my beer, if that's a BIG difference I'll show you night and day differences and then apples and oranges.

  • 6 months ago
  • 1 point

I am saying is that the argument "most games only use two cores" is just random nonsense that is just being repeated ad nauseam.

And I'm saying that the idea of buying a 9900k and disabling HT for a slightly better overclock is not good general advice. It's a niche solution for some serious enthusiasts. And unless the topic is clearly in that vein it's a bit silly to mention, IMO.

You speak as if the 9900k is a dual core cpu with HT disabled. Did you forget about the other 6 cores?

I'm not sure how you inferred that. At any rate my argument is still going to be if you want a 9900k without HT, they already make one, it's called the 9700k and it's cheaper. And anyone who really wants to buy a 9900k for the slightly better binning so they can disable HT and OC slightly better is free to do that, and is probably decided on that course and and isn't going to be asking CPU opinions, or looking for a dialogue about it. And hopefully understands that their decision doesn't make sense for a majority of users.

4.7ms vs 9ms is a BIG difference on a 1ms gaming display.

Well... you can believe that. And there's a million things I could say about this. Like that 1ms value is marketing nonsense and in no way represents real-world response times which are going to be quite a bit slower. So either you don't know how that number is bogus, or you're hoping I don't know that it is. Either way it makes your claims suspect. Besides claims in general centering around single digit ms latencies being noticeable by a human being are suspect. If you're a believer I'm not going to convince you otherwise, and you're just going to assume anyone who can't see a BIG difference isn't as informed, aware or isn't as sensitive to your BIG differences as you are.

Sort

add arrow-down arrow-left arrow-right arrow-up authorcheckmark clipboard combo comment delete discord dots drag-handle dropdown-arrow errorfacebook history inbox instagram issuelink lock markup-bbcode markup-html markup-pcpp markup-cyclingbuilder markup-plain-text markup-reddit menu pin radio-button save search settings share star-empty star-full star-half switch successtag twitch twitter user warningwattage weight youtube