9800X3D vs ??????

[Pocahawtness]

I am looking to buy a new PC. AMD is something new to me, so I am very unfamiliar with what they do.

The 9800X3D seems like a good choice, but I am a little concerned that it's being over-hyped. It seems that it's an E-Sports CPU, and not necessarily for me. I'm not sure about other AMD processors though. I know very little about them.

What are your thoughts on this? Is the 9800X3D a no-brainer for gamers, or are they really just for the 1080P players?

[Monk]

Originalmente postado por Administration:

Originalmente postado por A&A:

Yes. Here a few examples: The best benchmarks of:
Some random Xeon W9
https://novabench.com/result/7bdfde0b-820b-47d0-b313-cb0bd8d26f76

i9 14900k
https://novabench.com/result/db84178c-eb39-40dd-8faf-ac72640afe3f

Your i9 is twice as slow. Does it mean it's worse in gaming?

Even worse things can be seen if we open the database
https://novabench.com/parts/cpu
This ranking is very bad. It's even worse than Userbenchmark. Hell, they need to fix their website too, their time period filter is broken. So their comparison too and they want money to gain access to the full data?
How can you explain this
i9 12900 = i9 14900K = i9 10940X = Ryzen 9 3950X = i9 7940X

As I said, a higher result would suggest better performance.

Only 8n that test, look, stop trolling, you cannot believe the rubbish you are saying and I do not believe you think 1 synthetic benchmark has any relievence beyond itself, still it's a test I've never heard of nor can anyone figure out what it actually tests

[Administration]

Originalmente postado por Monk:

Originalmente postado por Administration:

As I said, a higher result would suggest better performance.

Only 8n that test, look, stop trolling, you cannot believe the rubbish you are saying and I do not believe you think 1 synthetic benchmark has any relievence beyond itself, still it's a test I've never heard of nor can anyone figure out what it actually tests

It does what a benchmark software does, it measures performance.

The higher the number, the better.

Novabench has been around for a very long time.

[r.linder]

Any "benchmark" that has results that place the i9-12900, i9-14900K, i9-10940X, R9 3950X, and i9-7940X all on equal footing is a benchmark with zero accuracy or relevancy. All of those CPUs perform wildly differently in reality, which means none of the benchmark results for any comparison can be considered remotely accurate.

UserBenchmark's also been around for a long time and it's also complete garbage, known for being so inaccurate that it's banned from many social groups and websites, and has become part of meme culture among computer hardware enthusiasts.

[Administration]

Originalmente postado por r.linder:

Any "benchmark" that has results that place the i9-12900, i9-14900K, i9-10940X, R9 3950X, and i9-7940X all on equal footing is a benchmark with zero accuracy or relevancy. All of those CPUs perform wildly differently in reality, which means none of the benchmark results for any comparison can be considered remotely accurate.

UserBenchmark's also been around for a long time and it's also complete garbage, known for being so inaccurate that it's banned from many social groups and websites, and has become part of meme culture among computer hardware enthusiasts.

novabench and userbench are 2 entirely different software companies. novabench is trusted by millions and with a positive reputation according to google.

its an awesome feeling when someone tells me to use their software to benchmark my pc with and I absolutely destroy their score result.

if novabench is showing that the 14900K has 2000 to 4000 more gFlops over the 9800X3D on average, im going with the 14900K. its clearly the stronger cpu. the king.

[A&A]

Originalmente postado por Administration:

novabench and userbench are 2 entirely different software companies. novabench is trusted by millions and with a positive reputation according to google.

its an awesome feeling when someone tells me to use their software to benchmark my pc with and I absolutely destroy their score result.

if novabench is showing that the 14900K has 2000 to 4000 more gFlops over the 9800X3D on average, im going with the 14900K. its clearly the stronger cpu. the king.

Meanwhile i9 14900K has 16 E cores that are equal like 6-8P cores. The average Cinebench like benchmark then.
The Ryzen has non of these cores, but it doesn't need them anyway.

The i9: 4011/14=286.5
The X3D: 2577/8=322.125 (Faster cores? No problem)

btw the Ryzen 9 9950X is the fastest desktop CPU acording to your benchmark and you turned a blind eye?

[Illusion of Progress]

Ah, yes, because GFlops alone is always the single limiting factor that determines performance. That why real world results always mirror the hierarchy this particular synthetic constructs. Oh, wait...

[Administration]

Originalmente postado por Illusion of Progress:

Ah, yes, because GFlops alone is always the single limiting factor that determines performance. That why real world results always mirror the hierarchy this particular synthetic constructs. Oh, wait...

novabench doesnt just measure cpu performance. novabench also measures performance from other areas of the system such as the gpu, memory, and storage.

for example, my m2 storage reads at 21000MB a second in crystal disk and that is the same amount novabench reported my reads at so there is no question whether novabench is accurate or not.

[Monk]

just that it is irreleant when looking at gaming performance.

atleast use a 3dmark bench if you arent going to use actual games.

[Administration]

Originalmente postado por Monk:

just that it is irreleant when looking at gaming performance.

atleast use a 3dmark bench if you arent going to use actual games.

i get an excellent score on 3dmark as well and its likewise to novabench. ;)

[PaulKrawitz]

I feel like the 7950X3D is better. 🤷🏼‍♀️

[Illusion of Progress]

Originalmente postado por Administration:

Originalmente postado por Illusion of Progress:

Ah, yes, because GFlops alone is always the single limiting factor that determines performance. That why real world results always mirror the hierarchy this particular synthetic constructs. Oh, wait...

novabench doesnt just measure cpu performance. novabench also measures performance from other areas of the system such as the gpu, memory, and storage.

for example, my m2 storage reads at 21000MB a second in crystal disk and that is the same amount novabench reported my reads at so there is no question whether novabench is accurate or not.

So what does a piece of software doing something else have to do with me saying that a single measure of CPU performance isn't always the limiting factor for additional performance in reality?

If you're gullible enough to think "this one synthetic matches up with this other synthetic, therefore it's always accurate to all software in reality", then I've got oceanfront property in Switzerland I'd like to sell to you.

Instead of distracting yourself with other synthetics, perhaps consider I was precisely trying to highlight that synthetics do not always match up with real world performance.

[Administration]

Originalmente postado por Illusion of Progress:

Originalmente postado por Administration:

novabench doesnt just measure cpu performance. novabench also measures performance from other areas of the system such as the gpu, memory, and storage.

for example, my m2 storage reads at 21000MB a second in crystal disk and that is the same amount novabench reported my reads at so there is no question whether novabench is accurate or not.

So what does a piece of software doing something else have to do with me saying that a single measure of CPU performance isn't always the limiting factor for additional performance in reality?

If you're gullible enough to think "this one synthetic matches up with this other synthetic, therefore it's always accurate to all software in reality", then I've got oceanfront property in Switzerland I'd like to sell to you.

Instead of distracting yourself with other synthetics, perhaps consider I was precisely trying to highlight that synthetics do not always match up with real world performance.

it's not that complicated.

if my pc scores higher than someone elses, it's because my pc is obviously better and faster than theirs.

[r.linder]

Originalmente postado por Administration:

Originalmente postado por r.linder:

Any "benchmark" that has results that place the i9-12900, i9-14900K, i9-10940X, R9 3950X, and i9-7940X all on equal footing is a benchmark with zero accuracy or relevancy. All of those CPUs perform wildly differently in reality, which means none of the benchmark results for any comparison can be considered remotely accurate.

UserBenchmark's also been around for a long time and it's also complete garbage, known for being so inaccurate that it's banned from many social groups and websites, and has become part of meme culture among computer hardware enthusiasts.

novabench and userbench are 2 entirely different software companies. novabench is trusted by millions and with a positive reputation according to google.

its an awesome feeling when someone tells me to use their software to benchmark my pc with and I absolutely destroy their score result.

if novabench is showing that the 14900K has 2000 to 4000 more gFlops over the 9800X3D on average, im going with the 14900K. its clearly the stronger cpu. the king.

"According to Google" means nothing, UserBenchmark pays Google to be at the top of search results and continues to allow it even though the popular consensus is that it's garbage.

So what Google says is or isn't good is generally BS.

[r.linder]

Originalmente postado por Administration:

Originalmente postado por Illusion of Progress:

So what does a piece of software doing something else have to do with me saying that a single measure of CPU performance isn't always the limiting factor for additional performance in reality?

If you're gullible enough to think "this one synthetic matches up with this other synthetic, therefore it's always accurate to all software in reality", then I've got oceanfront property in Switzerland I'd like to sell to you.

Instead of distracting yourself with other synthetics, perhaps consider I was precisely trying to highlight that synthetics do not always match up with real world performance.

it's not that complicated.

if my pc scores higher than someone elses, it's because my pc is obviously better and faster than theirs.

Good for you, too bad nobody really cares what your benchmark results say because they mean nothing in reality where you haven't been for awhile. While you're wasting time worrying about worthless benchmark results that are very obviously skewed, people are just using their machines and enjoying life.

[Illusion of Progress]

Originalmente postado por Administration:

it's not that complicated.

if my pc scores higher than someone elses, it's because my pc is obviously better and faster than theirs.

I'm interpreting this as "I don't respect what is a pretty complicated subject, so I tell myself it's simple and boil it down a singular number and tell myself that this single number is always absolute to have any chance of understanding it."

You've been corrected enough in this thread alone (which has seemingly run its course and at this point you've turned it into arguing about whether your own hardware is the best), so you're either willfully choosing to remain wrong, or you know better and you're trolling. In either case, good luck with that going forward, but if it's the former, then I hope you learn better one day for your own sake.