Best CPUs (Archive)

Status
Not open for further replies.

NoUntakenNames

Prominent
Mar 31, 2017
3
0
510
The 1600X doesn't deserve to be on the list, because it adds very little over a slightly OC'd 1600. You're tossing a bundled cooler and $30 for 400 Mhz. I'd recomend a 1600 and a $30 cooler like the H7 or Hyper 212, which would allow for an OC higher than 3.6 Ghz and a few bucks in your pocket as you don't need to spend extra money for an "X".

Otherwise, the updated list is very welcome, and I'm really happy to hear that Ryzen has improved for gaming. Good work, guys.
 

Teddie_Bear

Commendable
Dec 6, 2016
2
0
1,510
The 1600X has 6 physical and 12 logical cores, and not "six logical cores and 12 threads", since logical cores is the amount of physical cores multiplied by the amount of threads that each core can run, essentially the same as the amount of threads.
 
I still find the recommendation of the Core i3-7100 strange, considering it's far more expensive than a Pentium and pretty much the same performance. But otherwise this is a pretty nice list of recommendations.
 

FritzEiv

Honorable
Dec 9, 2013
253
0
10,780


Paul's working on something along these lines. We keep getting bombarded (in a good way, of course) with more SKUs to test, but trust that we have some things in the works here, so stay tuned.

 

Supahos

Expert
Ambassador
No, changing resolution has nearly no impact on the CPU usage. Actually assuming you aren't cpu bottlnecked at say 1440p with a setup dropping to 1080p will actually increase CPU load as the GPU can handle more frames. 60 fps at 720-1080-1440p all require nearly the same CPU load. Lock frames to 60 and check for yourself
 

rasmusdf

Distinguished
Mar 25, 2011
14
1
18,515
Nice set of recommendations. It's great there are some clear alternatives to choose from. It will be interesting to see what AMD can bring to the budget space.
 

namad7

Reputable
Jul 2, 2017
6
2
4,515
@elbert I think the ryzen 7 1700 was left out because 8 cores is too many for gaming, wasting heat that isn't useful and lowering clockspeeds. Which is fine for the 1700x or i7-7700k because the unlocked multiplier means at the very top end you can just liquid cool them and get the cores AND base clockspeed both. I have no idea who the ryzen 7 1700 is for though, excepting streamers, or other office rendering tasks. Then I think the ryzen 1700x was left off for being arbitrarily above the maximum price they ever recommend. Basically the top of the line pick cannot have the locked clockspeed of the ryzen 7 1700, maybe the ryzen 7 1700 deserved to be on the list, but only if they added another price point between the ones they already had.
 


The Ryzen 7 1700 is fully unlocked for overclocking, and will generally hit just about the same overclocks as the more expensive 1700X and 1800X.
 
What's with the cpu pricing on the 7600k? The Amazon price says $350 by the auto price grabber (at least I assume that's what it is). This article was written yesterday yet today I'm looking at Amazon and the price is $239.99 in stock. Are the price to performance charts also using this oddly high price of around 45% inflation over what it's actually available for?

Nevermind, I see what happened. It's overpriced since it's a bundle deal for a 240mm aio cooler. There are less expensive bundle deals on the same page. Not sure why we're comparing an outrageously high price of the i5 paired with a high end liquid cooler. Yes the 7600k doesn't come with a stock cooler but neither does the 1600x and it's listed and priced at cpu only. What's the price of a 1600x if the user decides to go full EK custom loop at around $450?

A bit drastic but nothing mentioned ahead of the link, instead it's a poor misrepresentation. For those who don't know any better they'll see the i5 as way out of their price range when in fact it's identically priced to the ryzen 1600x. A more fair comparison of price since both are sans cooler. Or someone gets the impression the i7 7700k is cheaper and then realizes it also needs a cooler like the 1600x and 7600k.
 


The Amazon price links are dynamic, so they can do silly things. The price/performance charts are based on realistic prices at the time the article was written.
 

woodscrews

Honorable
Jan 14, 2013
81
0
10,630
the 99th percentile results dont tell all though, this test should also have included results for the worst 1% of frames too, because it seems one thing that the 7700ks monstrous single threaded performance cannot deal with is the low minimums vs almost all high core count chips.
 


The 99th percentile frame time is the frame time that 99% of frames are delivered at or under. So it's the level separating the 1% of worst frames from the other 99%. It's a more robust figure than "minimum framerate" that was used in the past. It gives a more representative view of stutters, by ignoring the most extreme outliers.
 

SR TEE

Distinguished
Jun 13, 2013
35
3
18,545
I'm glad to see Tom's Hardware addressed the concerns of it's readers by actually factoring in motherboard and H/F costs giving much better picture of dollar to performance ratio.
 
Status
Not open for further replies.