I said £50 give or take, so that is £220.
Fair enough.
Secondly you are living in the previous world where there was no CPU competition, and complete stagnation from 2010 until 2017, that isn't the world now, and in 5 years it will be totally different again, look what has happened in less than 4 years, vs. the previous 7.
As I said I've not been keeping up with hardware developments for a few years. What exactly has happened in the last 4 years that is different to how things worked before? I know that the CPU power required by games has only increased very slowly in the last few years, in terms of raw CPU power my OCed 3770K probably is still enough for most games. I'm mainly upgrading to get access to DDR4 and M2 drives.
How is spending £270 by selling a part you already own (which will be worth £80 if you are lucky once the GPU shortage disappears), and topping up the budget by a tiny amount, not a good idea?
Yes I agree that selling one of my 980Ti cards now to top up my budget is probably a good idea.
Why spend more to gain little to nothing? Forcing your self to spend 2-3x the amount, in the false sense it will last longer is just that, false. You think that by spending more to get an expensive board that offers an extra couple of USB ports, and a beefed up VRM is going to make the system perform better? Or almost doubling the cost of the CPU 10400f, to 11600K then sticking an £100 AIO on it, that will last longer? It won't.
This is where you're losing me. How can buying more powerful hardware NOT make it last longer (i.e. maintain decent performance in latest games)? If I buy a B560 board now, I'm going to need to upgrade that in 2-3 years. If I buy a Z590 board now, it will cost marginally more but potentially last nearly twice as long before needing replacement. So therefore I save money over the 5 years by buying a Z590 now compared to a B560 now and another mobo in 2-3 years.
See the example below, a 10 game benchmark at 1080p ultra, using an RTX 3090. The 11400f (10400f is the same average almost) is behind the 11600K by a fraction, if you really want to get the better part and waste money get the 5600x, or if you want it to last longer get an 8c/16t part, like the 5800x.
Thanks for these graphs, I wasn't aware how close the 10400 and the 11600 are in real world gaming performance, that's pretty surprising given the difference in specs and single core performance. I know up to a few years ago, despite a majority of games/applications happily using 2 or 4 cores, single core performance was still a strong predictor of overall gaming performance of the CPU, is that not the case anymore?
So now I'm seeing the savings benefit to buying a 10400 now, and upgrading to a used 11600K or even 11900K in 2-3 years because I can do that without the need to upgrade other components at the same time.
You've obviously got your heart and head set on something, and don't want advice, just want confirmation.
There's no need to be rude. I am happy to change my component choices if I am sufficiently convinced it's worth it to do so. I am the type of person that needs to understand the 'why' (in great detail) behind any advice I am given before I can even think about following that advice.
It's also that what you're saying about doing a minor upgrade now and another in 2-3 years goes against all the experience I have with DIY projects in various hobbies. Spending a bit more for a component that will last longer before needing replacement/upgrade has always been a strategy that has served me well. Why, in this specific situation, is this not the best value for money strategy when it seems to work everywhere else?
I get that a more expensive component might not give much extra performance over a cheaper component
right now, but that's not the point. The PC requirements of games are steadily increasing all the time, therefore a more powerful component will require replacement/upgrade (to provide acceptable performance in the latest games at that time)
later than a less powerful component. What exactly are you saying is wrong about this statement?
What ever you buy unfortunately is already outdated with AMD doing AM5 next year with a bigger jump in performance over Zen 3, then Zen 3 was to 2.
And intel releasing LGA 1700 Q3 (stores Q4)
Intel is every 2 Gens they change socket , AMD is currently 3 (excluding Zen+ and potential Zen3+) . Though both new sockets will run DDR5 which is costly !
So you're suggesting I do nothing, sit on my thumbs and save more money until LGA 1700 and DDR5 hit the market? Tbh from some of the advice I've received here that option is looking more and more sensible all the time. Then again there's always something new around the corner and I can't wait forever.
With intel, recommend getting iGPU incase you have a GPU RMA or situation where you don't have a GPU. Major advantage over Ryzen currently and it's not till Zen4/ryzen 7000 it comes as standard.
Certainly very sound advice, but I've already got that covered, I have a GTX 460 lying around somewhere just for that reason. Any old parts that are not worth selling get stored away, I probably have enough spare parts for one complete DDR2-era low-end PC.
If I were you I'd sell those 980s whilst your can get good money for them and just buy a complete new system as others have suggested but it's your call!
Yeah I'm almost certainly going to be selling one of them since whatever upgrade I make will not support SLI. Since I only paid £150 each for them I will actually make a small profit selling one.
Yes, should be fine
Seems reasonable yes.
11600 - Normal chip,
...K - higher power used by default for more performance, allows overclocking,
...F (KF is K + F) no IGP (a mini GPU in the CPU), these tend to be a little cheaper.
Nobody knows. Currently 2080 ti's, 1080 ti's and else are selling around their MSRP (original price) despite being used and several years old. There's no reason to believe that this won't still be the case in the future.
No, NVME SSD's tend to only be about 15% faster than 500MB/s SATA SSD's once you take CPU bottle necks and random access into account. Probably still worth getting though as PCIE4 can enable upcoming features like smart memory.
32 GB is more than enough, most users won't hit 14 GB ever. The effects of RAM vary based on the processor and task at hand, but the speed delta between best and worst RAM tends to be at most 5%. Between 3600-3200 you're unlikely to notice any difference especially as higher clock speed RAM tends to have more latency (eating up some of the performance benefit). Either 3200 or 3600 should be unnoticable.
Thanks so much for answering my questions directly and itemising the answers. That post was incredibly helpful.
One question though, how are most users getting by with only 16GB? Some of my games are using 10GB, discord takes a few hundred meg, windows itself constantly uses around 2GB, then if I want to alt+tab to look something up on the wiki for the game or youtube that's another GB or two for the browser. On some of my larger games I cannot open a browser at the same time or I'll run out of RAM/VRAM and the game will crash to desktop so I have to use my laptop instead. Am I doing something wrong haha?
