Intel & AMD Processor Hierarchy

Page 10 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
sycoreaper said:
Thats the same justification HWUnboxed uses but that really only matters for synthetic means. Actual numbers would be equally of use.
Click to expand...
You asked a question. Not only did you get an answer, but you already were aware of the answer.
Now you're making it clear that you don't want the answer, but instead you want an explanation that YOU want, rather than reality.
 
King_V said:
You asked a question. Not only did you get an answer, but you already were aware of the answer.
Now you're making it clear that you don't want the answer, but instead you want an explanation that YOU want, rather than reality.
Click to expand...
Whatever mister hostility, you approached me. I'm free to give my opinion. If you don't like it, move along.

Not wasting my time reading any responses of yours so don't bother writing one.
 
M42 said:
It's a common theme in almost every area. Take high-end stereo equipment for example. You can spend 10x the money for 5% improvement. People buy high-end usually because they can. To some people, none of the highest-performing CPUs are worth the price. To others, price does not matter. For those people, the fastest gaming CPU is not actually the 5800X3D.
Click to expand...
Actually at some salary income per hour it makes sense to just buy the best CPU on the market. If your buying decision usually takes 12 hours for the price-optimized cpu over a three year product-life period, then if you earn $50/hr, you can afford the extra $600 for the very best CPU and RAM and mobo if you skip the 12hr CPU selection process and just buy the best, saving 12 hours of researching, googling, looking at articles, reading forums, etc. With the best product it probably lasts 3.5 or 4 years so you'd more than break even ...
 
Last edited:
There seems to be an error in "2023 Multi-Threaded CPU Benchmarks". It contains two 7950x3d entries, the second one listed as 12 core. Also 7900x3d isn't listed, but the wrong 7950x3d entry can't be a misnamed 7900x3d entry because it's got a higher score than 7900x.
 
cangelini said:
Here is a resource to help you judge if a CPU is a reasonable value: The gaming CPU hierarchy chart groups CPUs with similar performance.

Intel & AMD Processor Hierarchy : Read more
Click to expand...
Wonderful resource, but I was really disappointed to see that the i3-12100 (or 12100F) is missing from these charts. Especially since the i3-10100, and i3-8100, for instance, are present...I have an i3-12100F, thanks to your article "Little Gaming Giant", and I would really like to see how it stacks up to others.

Still, it is wonderful as is. Thanks.
 
Both the i3-12100 and i3-13100F are represented in the charts. Third to last row and fourth to last row. Also appears on the main charts and single threaded charts.

They are missing on some of the oldest charts, but the relative performance compared to their contemporaries is certainly there.
 
LoL i did not know a 78000X3D existed and costing 669 $
Maybe i am crazy but it does not compute, fix the wrong price and make it an existing cpu
 
cangelini said:
Here is a resource to help you judge if a CPU is a reasonable value: The gaming CPU hierarchy chart groups CPUs with similar performance.

Intel & AMD Processor Hierarchy : Read more
Click to expand...
Toms Hardware a source for Windows users only. Too bad. AMD and Linux have been a constant presence in High Performance computing. Always overlooked by Toms Hardware always will. Toms will not invest in Personnel who are Linux and High Performance Computing knowledgeable. Many single users are now build HPC single systems and enjoying development in many areas of High Performance computing. I would like to see Toms venture into Linpacs, and performance monitoring to include Networking and Message Passing Interfaces. I understand it will take time. You can get into really interisting subjects like module tuning both bound within the kernel and outside of the kernel.
 
TomsHardware, you might want to look at the second set of graphs (Windows 10 ) on page 2. It seems as if either the bars for Team Blue and Team Red are mixed up, or the bars shifted up one, and now no longer algin with the corresponding CPU. Example: Intel CPU's next to a red (AMD color) performance bar.
 
  • Like
Reactions: Order 66
How is it possible that in newer test Intel Core i9 13900K scores 162 (first chart Average FPS (Geomean) Entire Test Suite 1080p Windows 11) while in older test (2nd page of article, first chart) it scores 212? 50fps less new vs old seems a lot of difference to me (considered that header of graph says exactly the same).
Other CPU present in both graphs show similar big difference of fps
 
  • Like
Reactions: Order 66
Emiloz81 said:
How is it possible that in newer test Intel Core i9 13900K scores 162 (first chart Average FPS (Geomean) Entire Test Suite 1080p Windows 11) while in older test (2nd page of article, first chart) it scores 212? 50fps less new vs old seems a lot of difference to me (considered that header of graph says exactly the same).
Other CPU present in both graphs show similar big difference of fps
Click to expand...
Difference between 99th percentile and full average I would assume.

Also the non-overclocked result was 200.
 
  • Like
Reactions: Order 66
Yarberger1 said:
I have an i5-3570. I have previously been told that an upgrade to (for example) an i7-8700K would see a significant improvement and would be well worth the investment. This article seems to say the exact opposite.

I presently have a 1920x1080 monitor and GTX960 GPU. My system is used for general office applications, surfing and limited gaming (WoW, with an intention to switch to Elite: Dangerous shortly).
Click to expand...
I replaced my Ryzen 5 with an i5 3340 as I just use my pc for web design. I am planning to go big later down the line but for now i am good. For me it goes about use and price to performance. If I wanted to build my own new system now, The specs I want will cost me over 15k South African Rand. I got this I5 with 16gb ram which can oc to 2133mhz and rx470 for around R1500 in my currency. Obviously the Ryzen will kill it but not on a price to performance bases.
 
I truly appreciate these kinds of consistently-updated metrics. I have long been an AMD fan, and it's time to replace my Ryzan 7 machine. Does anyone have any experience running WSL2 on Windows 11 on AMD chips? My older machine had occasional WSL crashes, but that was a chip from 2018 and Windows 10. Thank you in advance to anyone who's had any experience that might shed some light for me.
 
Yarberger1 said:
I have an i5-3570. I have previously been told that an upgrade to (for example) an i7-8700K would see a significant improvement and would be well worth the investment. This article seems to say the exact opposite.

I presently have a 1920x1080 monitor and GTX960 GPU. My system is used for general office applications, surfing and limited gaming (WoW, with an intention to switch to Elite: Dangerous shortly).
Click to expand...
I'd suggest to be sure run your PC thru the Detection process here, and them you can check most any game. One caveat occasionally it will tell you it won't run a game, but if you look at it's rating for each component, they all equal or exceed the requirement, so it might run in thaat case. It will also rate the Recommended Reqs as well as the minimum.

https://www.systemrequirementslab.com/cyri/requirements/elite-dangerous/20402
 
Is there a reason why the raptor lake Intel chips are running OC: XMP DDR5-6800 & AMD Zen 4 chips running only OC/PBO: DDR5-6000 ?
Also, its not indicated what FCLK the Zen 4 chips are running in these tests. This is a metric that can influence performance outcomes for these chips.
 
Hotrod2go said:
Is there a reason why the raptor lake Intel chips are running OC: XMP DDR5-6800 & AMD Zen 4 chips running only OC/PBO: DDR5-6000 ?
Also, its not indicated what FCLK the Zen 4 chips are running in these tests. This is a metric that can influence performance outcomes for these chips.
Click to expand...
I can't answer the bit about FCLK, but the two platforms are running different RAM kits. The Intel has G.Skill Trident Z5 RGB DDR5-6800, while the AMD has G.Skill Trident Z5 Neo DDR5-6000. I'm pretty sure AMD (and likely Intel as well) sent those kits with the review processors.

The important aspect is that Z5 Neo has AMD EXPO support that's supposed to be more compatible and faster than kits that only have XMP profiles, at least for AMD platforms. I know at some point I saw that someone had tested XMP vs EXPO performance with Zen 4 processors and found that "slower" EXPO memory generally outperformed "faster" XMP memory, at least up to a certain point.
 
  • Like
Reactions: King_V and Hotrod2go
King_V said:
In addition to that, while I'm not sure if my memory (ha!) is serving me well at the moment. I seem to recall seeing that 6000 was considered the "sweet spot" for Zen 4, wasn't it?
Click to expand...
Yes, 6000 was deemed the so called "sweet spot" with Zen 4 when OC when in 1:1. However there are mitigating factors affecting that. FCLK is one & minor RAM timings have an impact on Zen 4 performance more than Intel platforms. But that's getting complicated for most over clockers. There are sites out there that discuss this in more depth for them if they search for them.
It's also debatable if one CCD chips benefit from 6000 more than say 8000 (2:1), the general consensus seems to be that two CCD chips perform better with 8000 but like all things, depends on the benchmark &/or game engine.
 
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom