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

Used 8700k or i5-9600k?

caparts
  • 17 months ago

I plan on building a 2nd system (already got a R7 2700 - purchased on sale - saved over $100 on the 2700X - reason why I went with it).

I plan on getting a 'cheap' Intel system - probably itx but looking at the processors - I have my eye on a used 8700K ($370 CAD) but noticed the 9th gen i5-9600K is around $350 retail right now. Does it make sense to stick with the i7?

I think single thread performance is very similar/comparable which makes the 9th gen i5 cpu a compelling buy in comparison but the multi-threaded performance is much lower (no hyperthreading is the reason?). There is solder with the 9th gen i5 whereas the 8700k would probably have to be delidded to have the option of really overclocking, right?

Which is the better choice? Just curious. The i5 is just under $400, the used i7 is $370 and for reference, I can even get an 8086k for $450 but I think that's over doing it - more $$ for not much more performance than the i7-8700K and it still doesn't have solder (although, it's better binned but extra $50 is not worth it - it's 'used' but meaning someone won it but is selling it).

Please advise (note: it will be used as a hackitosh but also because I use Linux/Windows and would rather have the 2nd computer and not have to reboot all the time when switching).

Comments

  • 17 months ago
  • 3 points

You don't need to delid the i7 for overclocking, thats just an added bonus to pushing the chips performance a little higher with lower temps. With a decent cooler you can easily push the chip around the 5Ghz mark without delidding.

For $20 more, the i7 is definitely the better buy, thats 6 cores and a 12-t hyperthread count hence 30-40% more performance for tasks which are multi-core+HT optimised.

The 9600k is basically the 8600K (both absent of HT). Difference being, the solder between the IHS/DIE and a slightly increased clock rate out of the box. With all that in the mix, the 9600K remains at par with the i7's single core frequency and yet with the i7's higher thread count for only $20 - it's a clear winner and definitely the better future proofer.

The question you'll need to look into further: Can a hackintosh build benefit from hyperthreading without inconsistencies or drawbacks. Personally, I would make a new thread to ask some of the more experienced users who are familiar with hackintosh makeovers

  • 17 months ago
  • -3 points

Delidding is damn near required for overclocking the 8700k. It’s toasty on stock settings, pumping more voltage just causes more issues.

Unless you enjoy running it at like 90C

Also the 9600k can’t be compared to a 8600k. The die changed so it’s a little more than “basically the same”

If he has issues running Hackintosh, he can always disable hyper threading

  • 17 months ago
  • 1 point

Overclocking on ITX can be tricky due to cooling, so be careful with that. A liquid AIO should handle either of these well, but air cooling is iffy at best.

If you're confident in your cooling solution, go for the 8700k. Using Userbench as a reference, the single core speed is virtually identical between these two, and the 8700k is going to give you a much better multi-core speed. I'd say it's more than worth the price.

  • 17 months ago
  • 1 point

That's what I was thinking, too. But, I was wondering if there is any argument for getting the i5-9600k since it is pretty close in price (to the USED chip price I found) but a newer gen. chip. The multi-core performance is it's main disadvantage, I think - isn't that right?

  • 17 months ago
  • 1 point

Multiple core on a 9600k will be inherently worse than a 8700k

Since it doesn’t have hyper threading.

The tiny generational changes that we get on 14nm Intel are not enough to get that kind of performance

  • 17 months ago
  • 1 point

Okay. The 8700k it is, then. Thanks.

  • 17 months ago
  • 1 point

I've got a 7600K running at 5GHz with a 240MM AIO in an mITX system. Runs in the 60's when stress tested, 50's when gaming.

It's delidded though lol.

  • 17 months ago
  • 0 points

Userbenchmark is a **** reference

Use a review site that does proper benchmarks on the same system to reduce variables to just CPU

Also remember that hyper threading doesn’t truly double thread counts, it splits the physical core into two threads and makes it do more work (~30% more). It’s why CPUs with HT have crazy heat density normally.

But yeah I’d recommend an 8700k and a delid tool (or a razor and a steady hand) over a 9600k

  • 17 months ago
  • 4 points

Userbench is really good for vast amounts of easily-referenced data, pulled from real systems that haven't been optimized for benchmarks. There's some variables and user error, naturally, but it pulls from a very large data pool and gives you a good idea of what users are really getting from their systems. From an industry standpoint it's not reliable, but for the end user it's a great at-a-glance resource.

  • 17 months ago
  • 2 points

It's definitely not a **** reference lol.

It's not the end all of benchmark suites, but it is an easy way to see results of literally thousands of samples provided by regular people.

The "worst" part about it is the GPU testing suite, because it's DX9.

  • 17 months ago
  • 1 point

The issue is that I can limit my 8600k to 1c and 2.0ghz and it’ll treat the benchmark the same as if it was 6c @6.5ghz

Too many variables to be reliably used for comparisons

  • 17 months ago
  • 1 point

Both of those tests would be outliers and wouldn't effect the average results of sometimes tens of thousands of tests run by people who are either running stock or reasonable overclocks.

You're nitpicking. For that to be a real issue there would have to be thousands of people doing that.

If you really want to be anal about it, you need to see tests comparing the processors in the exact scenarios you will be using. Which isn't really feasible, at all. Benchmarks are always generalizations.

  • 17 months ago
  • 0 points

Good thing there are plenty of websites dedicated to benchmarking CPUs with plenty of different workloads. That the only variable changing is the CPU.

Benchmarks are generalizations, yes. But they reduce variables and give fair comparisons

Userbenchmark does not ignore outliners and you can weight results. You want accurate comparisons? Use proper benchmarks and properly compare them.

You do have tens of thousands of people running software in the background, etc that are resulting in inaccurate results. You want direct comparisons, not RNG results depending on your background processes

  • 17 months ago
  • 1 point

Well you guys answered my question without even asking :)

  • 17 months ago
  • 1 point

Where are you finding 8700K for ($370 CAD)?

  • 17 months ago
  • 1 point

Its $500.00 everywhere I look to $570.00 CAD.

[comment deleted by staff]
  • 17 months ago
  • 3 points

lol fools day is in April

  • 17 months ago
  • 2 points

the 8700k performs worse than the 9600k

quite the opposite https://cpu.userbenchmark.com/Compare/Intel-Core-i7-8700K-vs-Intel-Core-i5-9600K/3937vs4031

  • 17 months ago
  • 1 point

Really? Any benchmarks to support that statement?

[comment deleted by staff]
  • 17 months ago
  • 3 points

Lmao what?

  • 17 months ago
  • 3 points

I repeat, any benchmarks?

  • 17 months ago
  • 1 point

Worse at what? How much worse?

  • 17 months ago
  • 3 points

the 8700k is better, dont listen to him/her

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