[SOLVED] Icore 9600kf good enough?

Jul 29, 2020
13
0
10
I currently have a Rtx 2070. 16gb of ram and an icore 5 9600kf. I'm wanting to get more frames out of my games. Would an upgrade in cpu get me that? If so what CPU would be the one to get? I play games like Gta 5, black ops 3/4 and call of duty Modern Warfare
 

Tom_nerd

Reputable
Jul 15, 2019
107
13
4,615
personally I think that you have 2 options. firstly you could go for a CPU and motherboard change which would give you better performance but would also cost more. secondly you could just upgrade the CPU and not the motherboard. this all depends on your budget. what is it?
 
Jul 29, 2020
13
0
10
personally I think that you have 2 options. firstly you could go for a CPU and motherboard change which would give you better performance but would also cost more. secondly you could just upgrade the CPU and not the motherboard. this all depends on your budget. what is it?
£500 is my budget. Not that clued up on cpus of motherboards. And I play on 1920/1080.
 

Tom_nerd

Reputable
Jul 15, 2019
107
13
4,615
£500 is my budget.
luckily I have written a document for CPU`s under £500.

(I cant work out how to attach the word document so I am just going to copy and paste it)
(I also put in some explainers to make a more informed desicion)

CPU review
I am writing this on the 9th of July. The reason I say this is that 2 days ago the new XT CPU’s from AMD came out which I will get onto shortly. So, let’s get on with the document. In this document I will list what CPU’s I would recommend for under £500. Firstly, I am going to be doing some explainers (I could not think of a better name). so that you will be fully prepared to read the document and hopefully everything makes sense.

So, for the first explainer, I am going to assess the argument of cores versus clock speed which is a question that has troubled many gamers. Normally the answer to that question is that it depends, which in their defines is technically not wrong but, that person who said that is really annoying because it has left the person who asked the question even more confused. It basically goes down to how the developer wrote the game although there are some things to tell how many cores it uses; the first way is to look it up online. There are loads of people on sites like Tom’s Hardware (no relation to me whatsoever) who have documented this stuff to inform people like that if that CPU will be overkill for that game. The next way to see how well optimised the game is to spread across multiple cores is looking at what API’s the game supports. API stands for Application Programming Interface, which is a bad name because it is way easier to understand than the name suggests. An API is basically how the game interacts with the hardware for example DirectX, which is an API written by Microsoft for Windows, has slowly adapted to the changes in the hardware industry as the versions have gone on. For example, if I was running DirectX 2 on a 16 core CPU, the program would only use 1 or 2 cores but, if you used DirectX 12 (the latest version) it may be able to use all 16 cores, its unlikely but it might, it is probably more likely that it would use 4-8 cores. There are loads of different API’s but for games there are 2 main different API’s: there is DirectX (as I mention earlier, it is created by Microsoft) and Vulcan (made by Khronos). Vulkan is the latest and greatest, but there are some advantages to DirectX as well. So, in an order of ranking those API’s its pretty simple because, it’s all the DirectX’s in history then Vulkan. Although you are probably not going to be choosing a game that will be able to take advantage of your new 64 core processor (there is no game at the time of writing that can use more than 18 cores, and that last bit about the 64 core processor was said in a sarcastic voice) your going to be picking a game for the actual content and experience of the game. So general guidelines for what types of games will scale in different ways. Games like city skylines or Tropico will generally scale quite well across CPU cores which is great because games like these also very heavily depend on the CPU instead of the GPU because there is so much going on in the scene, I will get onto what games use more CPU power and what games use more GPU power. For FPS games generally can scale across 4-8 cores, although generally in those kinds of games the GPU is the bottleneck. This means that developers will spend more time optimising the game for the GPU instead of finetuning the spreading of the workload across multiple cores. So, for those kinds of games having more CPU cores will not make much of a difference. For racing games, it generally uses around 6-8 cores depending on the game games that have a lot of things that are moving in them like character models will generally scale quite well across more cores.



Now for the second explainer, it’s about what kinds of games use different resources, for example a game like city skylines will use much more CPU power and is much more likely to be bottlenecked by the CPU. Because remember there will always be a bottleneck in your system because otherwise your system would be infinitely fast because nothing would be slowing it down, if you don’t get this then imagine this (if you do get it go to the red character) you have 10 runners who all have to stay as a group and have to run exactly parallel to each other, one of them will always have to be the slowest which means that all of the other runners have to slow down for that slow runner. This is the same as in your PC, if you have a 2080 TI then pair it with a core 2 duo, then the 2080 TI will have to slow down for the core 2 duo. In that analogy the core 2 duo would be Donald Trump and the 2080 TI would be Mo Farah. So now let’s get onto the RAW information! Generally games that have a lot going on, like having loads of different people moving will be much harder on the CPU whereas a game does not have much going on will be much easier on the CPU so it means that the GPU will be the slowest runner (that’s referring to the runner analogy earlier). Whereas games that run at a really high resolution and use very high-fidelity textures will be much more demanding on the GPU. A great video if you would like to do some extra research into it is this one:
View: https://youtu.be/dt5CNi0aEpI


So let me go over what the offerings are from AMD and intel from AMD it is the 3900X coming in a £440 and from intel it is the i7-10700K. there is also the 3900XT which is a very slightly higher performance than the 3900X although the BIOS support with motherboards is flakey at best. Which would mean that performance would be unstable. It also has not much of a performance increase over the 3900X as oppose to the jump from 3800X to 3800XT as they have a lot more performance headroom. For the i7-10700K it is basically the exact same performance as a i9-9900K, but it has a slightly higher power draw and has better compatibility with future motherboards. So now onto the individual comparisons.

Starting with the i7-10700K the main thing that is in its dis-advantage is the older process node. Even though this is one of the 3 “comet lake” processors commercially available at the moment the name “comet lake” is nothing but a name as intel has really been struggling to manufacture a new process node which means that they have gotten desperate and asked their marketing department to fix this. And they said that they would come up with a new name and make it sound like 10th gen processors are far superior to anything you have ever seen before. But in truth it is just a rebrand of an older CPU, the 9900K. although there are some advantages to going for a 10th generation intel CPU. Things like having WI-FI 6 built into the processor which means that for super high bandwidth situations it can give some advantage. Although generally speaking there are not many internet connections that can saturate the Wi-Fi 5 link. There is also having 2.5 gigabit networking which can be an advantage although, again that is only if your internet can use all that extra bandwidth. And you could get a PCIe network card for high network usage. Another downside to going intel is that at the moment you can’t get PCIe gen4 support. This means that PCIe 4.0 SSD’s will run at PCIe gen3 speeds which is half the speed. It also means that if future graphics cards start to use PCIe gen4 they will run at a slower speed on a PCIe slot.



Onto the 3900X. generally the main advantages of this CPU is its mullering of intel in workstation and general daily workloads like snappy internet browsing. For gaming AMD just beats intel but not by much. Although I think that AMD wins in so many other tasks that it means that the choice to pick AMD is much easier to do. It may appear that AMDs option appears to cost more than intel’s. but actually because you would have to buy a cooler as well. Another reason that AMD costs less is that the motherboards cost around £200 more than AMDs. Some of the extra reasons I will get onto in the graph.


I am going to go through each column and compare the 2 CPU’s in the category. Starting with the main differentiator between CPU’s, the price. They are only £10 apart and still under the £500 price point so I ignored the price difference. Now onto the second column. The 3900X has 4 more cores which can make a very large difference for tasks that can spread across all those cores, which now are only for workstation tasks. Hopefully in the future games should be able to scale cross more cores. An example of this are games that have come out in the last couple of years they have been scaling across around 8 cores and if things keep improving the way they are, we should be using 64 core CPU’s by 2030. Now onto the clock speeds, the main thing to consider when looking at these 2 clock speeds is that they are built on different architectures. So, if we excluded all other variables. AMD will be twice the actual clock speed compared to intel so 4.6GHz would be 9.2GHz. this is because AMD is built on the 7nm process node whereas intel bases their process node on the 14nm process which is twice as large so therefore twice as inefficient. Although there are loads of other variables which mean that it is around 1.25 times more in-efficient so 4.6GHz should be around 5.75 GHz. Now onto the amount of L2 (level 2) cache, level 2 cache is the kind of thing that can help game performance and because AMD has twice as much L2 cache which is a massive point in AMD’s favour. Now onto L3 cache this is mostly used in workstation tasks and that is why AMD thrashes intel at workstation tasks because it has 4 times more L3 cache. Now onto the TDP (Thermal Design Power) it may seem that intel loses but that is not the case intel gets absolutely annihilated by AMD. This is because AMD measures their temperatures at their boost clock whereas intel measures it at their base. This means that the 125-watt TDP is more like 145 watts. Another dis-advantage to intel in the heat department, is that AMD’s CPU comes with a cooler included whereas for Intel you must buy a cooler. This means that intel’s CPU costs more than AMDs. Now onto the socket, AMD uses the AMD socket AM4 whereas Intel uses LGA 1200. This means that AMD motherboards cost around £100-£150 whereas Intel motherboards cost around £250-£400. Now onto PCIe lanes, intel has 16 lanes which is only enough for a graphics card and nothing else like an PCIe SSD. Whereas AMD has 24 lanes which means that you could have a graphics card and 2 PCIe SSD’s. another thing to take into consideration is that AMD is using PCIe gen 4 which is twice the speed. This means that it would be 48 lanes to intel’s 16 lanes. They both have a limit of 128GB of RAM which I’m not going to address for multiple fairly obvious reasons. Now onto the ECC memory support, this means that you could use a 3900X in a “budget” server. Now back to the power consumption discussion, I have gone on about this earlier so I wont go on much more. But I just thought that I should mention that I got this data from a reputable source.
 
  • Like
Reactions: Samhaddow
Jul 29, 2020
13
0
10
What’s the full pc spec? What FPS are you actually getting?
I get about 70 fps on close to ultra on Modern Warfare and the same for GTA V.

CPU: Intel(R) Core™ i5-9600KF - 6-Core 3.70GHz, 4.60GHz Turbo - 9MB Cache

MEMORY: 16GB (2x8GB) DDR4/2400mhz Dual Channel Memory [+30] (Corsair Vengeance LPX w/Heat Spreader)

VIDEO: MSI GeForce(R) RTX 2070 8GB - Ray Tracing Technology, DX12(R), VR Ready, HDMI, DP, 4 Monitor Support [+60] (Single Card)

I also have a 1 terribite m. 2 drive with those games on
 
Are you currently on a Z390 mainboard, and, if so, what all-core clock speeds are being hit/maintained under load? (4.2-4.3 GHz would be 'standard', but an MCE -enabled mainboard or manually manipulation of BIOS would allow 4.6 GHz all-core or higher if power budget and thermals permit; many folks OC the 9600K variants to 4.9 or 5.0 GHz fairly routinely)

As the 9600KF has but 6 threads, and although gaming comparisons show it's min/average frame rates relatively high, any added loads/tasks such as live-streaming/ chat would indeed likely tank frame rates due to limited thread count, in which case I'd be looking for good deals on a 9900K or KF...
 
Jul 29, 2020
13
0
10
Are you currently on a Z390 mainboard, and, if so, what all-core clock speeds are being hit/maintained under load? (4.2-4.3 GHz would be 'standard', but an MCE -enabled mainboard or manually manipulation of BIOS would allow 4.6 GHz all-core or higher if power budget and thermals permit; many folks OC the 9600K variants to 4.9 or 5.0 GHz fairly routinely)

As the 9600KF has but 6 threads, and although gaming comparisons show it's min/average frame rates relatively high, any added loads/tasks such as live-streaming/ chat would indeed likely tank frame rates due to limited thread count, in which case I'd be looking for good deals on a 9900K or KF...
Asus Prime Z390-P: ATX w/ USB 3.1, SATA3, 2x M.2

It peaks at 4.3 ghz, I've not seen it go over that under any load.
 

TRENDING THREADS