• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

CLOSED

Every UE based game runs like dog shit because it’s a dog shit engine.
 
Sir Beregond said:
Fun chart. 4790k is what I had before this. :ROFLMAO:

Missing some specs though. What resolution is that data for? 1080p?
Click to expand...
Mr. Bluntman said:
I'd really like to know what detail settings were used for that chart as well...
Click to expand...

It's all in German, but from what I understand, this is on the launch day version of the game, with a 4090, 720p max settings, and it is seriously CPU limited on all of the CPU's in the test.

source

Educated guess based on my own testing is that even on that 9800x3d at 720p, at the 98.2fps average, the 4090 GPU is only loaded at just over 20% due to the CPU being unable to keep up.

On my Threadripper 3960x at 4k max settings without any fake frames or upscaling, I was GPU limited at about 51fps average. If I turned on upscaling that switched to a CPU limit at about 53fps average. So I decided to pass on playing it until I upgrade my CPU (soon). If I get a 5800x3d or 9950x3d I supposed I should be able to hit a mid 80's average if I use quality upscaling, from the testing I have seen. At that point I'll be GPU limited, but the CPU will be pushing at almost 90% of what it is capable of.

0.2% and 0.1% minimums will still be pretty bad, CPU limited at 44% and 35% respectively though. I guess there is no avoiding that.

The only criticism I have of these benchmarks is that they seem to use slower RAM than these platforms are capable of using, but I don't think that materially changes the outcome.

The game is pretty heavy on the GPU, but apart from relatively high VRAM use (you'll want more than 8GB VRAM even at 720p) it's just kind of "normal" GPU heavy. The CPU is where it gets extreme.
 
Last edited:
Zarathustra[H] said:
The game is extremely hard on the CPU, but it is well worth it.
Click to expand...
Again, if it runs like crap, I'd say it will bomb. However, we've certainly seen initial disasters become ... even .... good once "fixed". But no, if less than 1% of enthusiasts get an "ok" experience.... it will tank (until they fix it).

Edit: Remember what I said about the worthless pretty non-context graphs above? I mean, if the above is what you get at 8K Ultra + Smell-o-vision.... maybe "normal" stuff is quite reasonable game play wise.
 
I've said it in other threads, and proved it multiple times in games and benchmarks. If you are on AM4 Ryzen 5000, you can absolutely gain a lot of performance going with an All Core OC, CPPC off and SMT off. Some of this will depend on silicon lottery for sure, and motherboard and memory to some degree.

I originally did this a year or so ago, but i recently tested this all over again to verify it still applies today. This instantly raised my 1% lows on my 5950X and my 4090 is legit benching against people with 9950X3D chips and keeping up (and also beating alot of the time) in games and game benches.

Will it win in Cinebench? No, but i don't game in Cinebench... lol.
 
You don't even have to do an all core OC. Just max your power limits and set your curve to that. You get the grunt at the low end, and all the bald eagles to the cores on the top end.

My 5900X can do 260w with PBO.
 
freeagentt said:
You don't even have to do an all core OC. Just max your power limits and set your curve to that. You get the grunt at the low end, and all the bald eagles to the cores on the top end.

My 5900X can do 260w with PBO.
Click to expand...
So that doesn't perform as well as an All Core OC still. I found that the main issue is the boosting algorithm altogether and the latency it introduces into frame times and L3 cache. I'm not at my PC now, but I can post tomorrow. SMT on these chips was another culprit, although with less core CPUs, might be beneficial to still leave on if you need the threads.

Not only did games get a really nice boost, AIDA64 memory bandwidth and latency test showed why... I got an over 50% boost in L3 cache bandwidth. On my 5950X, it was the difference between 800~1000 GB/s vs. 1500GB/s. Solid, everytime. Ryzen 5xxx was very much memory cache starved, which is why the introduction of the X3D chips was a game changer (more memory close to processing). The chiplet design of the platform has always been a weak link. This configuration I did really helped a ton. If i decide to switch to a 9950X3D you can bet I'll still try to see if performance can be had in the same way... lol
 
III_Slyflyer_III said:
So that doesn't perform as well as an All Core OC still.
Click to expand...
It does pretty close, about 525GFlops static OC, which is 4500MHz, and 523GFlops with PBO boosts to 4550. So that is a worst case scenario load, and then it still boosts up to 5150 on a few cores. Only problem with that is the average effective clock will be nowhere near that. So you can run an OC of 4600-4700MHz static, but you will never be all load stable, and you will have to be careful about what you run, so it doesn't overheat, or crash lol.. that is no fun. The problem with PBO is that the majority of people do no know how to use it properly, and have shit cooling.
 
freeagentt said:
It does pretty close, about 525GFlops static OC, which is 4500MHz, and 523GFlops with PBO boosts to 4550. So that is a worst case scenario load, and then it still boosts up to 5150 on a few cores. Only problem with that is the average effective clock will be nowhere near that. So you can run an OC of 4600-4700MHz static, but you will never be all load stable, and you will have to be careful about what you run, so it doesn't overheat, or crash lol.. that is no fun. The problem with PBO is that the majority of people do no know how to use it properly, and have shit cooling.
Click to expand...
I agree on the PBO part and people not setting it up correctly, but you are also talking workloads. I'm talking strictly gaming. For gaming, the all-core OC route is proven to be a much better and smoother experience overall.
 
Best I could do mediocre stable was 4750, 4600 was fine, 4500 was good for all loads. I had no problem with PBO so I just mastered that. 4900 was not even remotely in the cards for an all core oc.
 
freeagentt said:
Best I could do mediocre stable was 4750, 4600 was fine, 4500 was good for all loads. I had no problem with PBO so I just mastered that. 4900 was not even remotely in the cards for an all core oc.
Click to expand...
I got 4.7Ghz Prime95 stable, 1.35V, L3 LLC. 4.8Ghz for me takes 1.3625V L2 LLC. I have a nice chip. Running 4.9Ghz gaming, takes 1.4V L2 LLC, but i can game and run cinebench, no issues, never crash, can run all day with it for what I do and it will not go past about 75C. P95 I don't know at 4.9, it hits 90C+ very fast even with 360mm AIO.
 
III_Slyflyer_III said:
I agree on the PBO part and people not setting it up correctly, but you are also talking workloads. I'm talking strictly gaming. For gaming, the all-core OC route is proven to be a much better and smoother experience overall.
Click to expand...
Good info but OP is building a video editing/rendering station, not a gaming computer.
 
Tsumi said:
Good info but OP is building a video editing/rendering station, not a gaming computer.
Click to expand...
Ahhhh... well, TBF, he did say "some gaming at 1440p"... the config would help with that 100%. 🙂 But yeah, I just get all excited trying to share how people can get even more out of their systems with a bit of tweaking and knowledge. Multiple use cases is why I have 3 profiles set in the bios, so that's always an option for people as well.
 
III_Slyflyer_III said:
Ahhhh... well, TBF, he did say "some gaming at 1440p"... the config would help with that 100%. 🙂 But yeah, I just get all excited trying to share how people can get even more out of their systems with a bit of tweaking and knowledge. Multiple use cases is why I have 3 profiles set in the bios, so that's always an option for people as well.
Click to expand...
Actually.... I thought I was responding to a different thread, my bad :facepalm:
 
I'm gonna have to hunt down a 9950x3d before these Taiwan tariffs from orange man take place. I was going to go to MicroCenter last weekend. Arrrgh
 
Whach said:
I'm gonna have to hunt down a 9950x3d before these Taiwan tariffs from orange man take place. I was going to go to MicroCenter last weekend. Arrrgh
Click to expand...
Looks like Taiwan has 64% tariffs on the US. I wonder if they'll keep them, or drop them?
1743657127045.png
 
WilyKit said:
No, but you live here and are complaining about reciprocal tariffs. Reciprocal means we have them because they do.
Click to expand...
trump can't spell reciprocal. He is putting tariffs on uninhabited islands. He is putting tariffs on because the countries have a VAT. They are in no way equivalent. That is OK though, A White House official, speaking to The New York Times on condition of anonymity, explained that Cuba, Belarus, North Korea, and Russia were not impacted by Trump's tariffs but a 10% tariff on Ukraine shows we are the nice guys.
 
I wonder how Canada will fare with ~100% tariffs on dairy? That's something that I found out just a few days ago when I bothered to look into tariffing a little bit. I noticed Mexico and Canada were not on the initial list. Either way, direct tariff calcs or based on "currency manipulation and trade barriers" (as the release showed on the top of the graphic) - it will be interesting to see if the US economy has a net benefit as it did with tariffs last time, or if most countries will raise tariffs. It seems there may be a few that are looking to FAFO (US is the largest consumer in the world), and several that have already dropped their tariffs (Israel and Argentina), and several that aren't responding as of yet (China and the UK are looking but seem neutral atm).

Time will tell.
 
Tsumi said:
Would be helpful if you can provide proof to the contrary.

But I'll do you one better: is straight from the horse's mouth good enough for you? Which was, by the way, also linked in the article if you bothered reading instead of being triggered by the source. https://ustr.gov/issue-areas/reciprocal-tariff-calculations
Click to expand...
GIGO, when you enter false data you get false answers. Anything provided by the trump administration is automatically to be distrusted.
 
narsbars said:
GIGO, when you enter false data you get false answers. Anything provided by the trump administration is automatically to be distrusted.
Click to expand...
Try anything from government should ALWAYS be distrusted. It will make you more savvy.
 
Tsumi said:
Would be helpful if you can provide proof to the contrary.

But I'll do you one better: is straight from the horse's mouth good enough for you? Which was, by the way, also linked in the article if you bothered reading instead of being triggered by the source. https://ustr.gov/issue-areas/reciprocal-tariff-calculations
Click to expand...
Not gonna do your homework for you. What good what it do anyway? It’s clear from the rhetoric we won’t convince each other. Vote for your interests and I’ll do the same.
 
Tsumi said:
Whether or not it's actually reciprocal seems dubious. https://www.cnn.com/2025/04/03/economy/reciprocal-tariff-math/index.html
Click to expand...

Agreed.

The "Tariffs charged to the U.S." on that chart is more like a random number generator than anything else, as it includes vague and difficult to enumerate things like "currency manipulations" and "trade barriers".

These things may or may not exist on a country by country, industry by industry and product by product basis, but they are WAY more complex than can be boiled down to a simple tariff equivalent percentage number, which means there is about a 100% chance that column is wrong.
 
You must log in or register to reply here.
Back
Top