Awesome Tips AMD says these are the same… We DISAGREE.
Remove your personal information from the web at and use code LTT for 20% off.
Remove your personal information from the web at and use code LTT for 20% off.
NexiGo is hosting product giveaways every month! Enter for your chance to win here: (Contest is open to anyone 13+ in the United States)
Do you really know if your CPU is performing the same as the ones we review? We don’t know. But we know that if we want to increase our testing capacity, we need to PARALLELIZE. But that means we need nearly identical test benches. And trying to make that happen sent us down a far deeper rabbit hole than we could have anticipated.
Testing Process Doc and Forum Post:
Buy an AMD Ryzen 7 7800X3D CPU: (Canada: ) (Micro Center: )
Buy a NVIDIA GeForce RTX 4909 FE Graphics Card:
Buy a GIGABYTE X670E AORUS XTREME Motherboard:
Buy G.Skill Trident Z5 NEO 6000MT/s CL30 2x16GB DDR5 RAM:
Buy a Samsung 980 Pro 2TB NVMe M.2 SSD:
Buy a Noctua NH-D15 CPU Cooler:
Buy a MSI MEG Ai1300P PSU:
Purchases made through some store links may provide some compensation to Linus Media Group.
Purchases made through some store links may provide some compensation to Linus Media Group.
► GET MERCH:
► GET EXCLUSIVE CONTENT ON FLOATPLANE:
► SPONSORS, AFFILIATES, AND PARTNERS:
► EQUIPMENT WE USE TO FILM LTT:
► OUR WAN PODCAST GEAR:
FOLLOW US
—————————————————
Twitter:
Facebook:
Instagram:
TikTok:
Twitch:
MUSIC CREDIT
—————————————————
Intro: Laszlo – Supernova
Video Link:
iTunes Download Link:
Artist Link:
Outro: Approaching Nirvana – Sugar High
Video Link:
Listen on Spotify:
Artist Link:
Intro animation by MBarek Abdelwassaa
Monitor And Keyboard by vadimmihalkevich / CC BY 4.0
Mechanical RGB Keyboard by BigBrotherECE / CC BY 4.0
Mouse Gamer free Model By Oscar Creativo / CC BY 4.0
CHAPTERS
—————————————————
0:00 Intro
2:25 Why Same Model ISN’T Same Performance
5:32 Sources of Variance
8:21 Gaming Results
10:43 CS:GO is Wonky
11:41 Gaming Results Cont.
12:39 Productivity Results
13:59 Our Selections
15:15 Testing Mobos and RAM
17:16 Final Discussion
21:26 Outro
Deals for Days. Big home savings are happening now.
CorsOla
Its just a fact and it has to be said guys….you are doing better than any government oversight could ever hope to accomplish. Stop think that government regulation could help with this, it will only result in a bureau that would get paid off to help them lie, cheat, and steal.
Does the performance difference follow the CPU from system to system? As in: did you swap the CPU to another test bench and the variance stayed the same?
I wonder if it'll be long before we see govt. oversight on this sort of thing. EU, you down?
Suggestion: Please can you keep the CPUs in the same order rather than sorting to Higher to Lower.
Doesn't this mean that the silicone lottery isn't dead?
only the variation from the promised baseline matters, so realistically the difference was 4-6% (unless they gave only the stats for the best one), which is an acceptable margin for a product that requires nanometer precision to manufacture
As an Australian, I appreciate the vegemite reference 😊
Please actually try to educate yourself as to why a regulation exists before suggesting similar regulation is also applicable to your 2% lows. Here's a fact: #1 You are statistically less likely to die as a result of a computer crash than you are to dies as a result of a vehicular crash. Please, in the future, don't compare apples to hand grenades.
Looks like you need Bayesian hierarchical models in order to progagate the sampling uncertainty correctly.
Clearly still so bitter about GN's video…
when recommending others sources they seemed to forget to include gamernexus. I wonder why hahahhahah
could it be a "faulty" motherboard, having many components on a motherboard would increase probability of "failure" compared to lets say a CPU or RAM
So when are you finally kicking userchmarkbench off the internet?
Finally some quality! Keep up the good work
you explained it perfectly and i never thought of it like that, just random chance what you buy has 10% less than supposed to
Ahhh finaly
Non engineering major can feel the pain of engineering major thesis
The Math wasn't mathing
How is this really useful? Interesting,but useful?!
Do you check for differences between test benches using one CPU before performing tests for multiple CPUs? Update: Nvm, its in video .:D
Why pokemon indeed…. 0_0 mmmmmm hey…. why indeed….. whoever could figure that one out >-< HAHAHAHA
Interesting that Gamer's Nexus didn't get a spot in the list of other creators.. Is Linus still salty?
Wouldn't it be a lot better to use cosine distance rather than euclidean distance? Euclidean distance in high dimensions tends to not work well because of the curse of dimensionality
If your problem is running windows, and you do not want to depart from microsoft, you can always upgrade to msdos. No more hidden processes running wild in the background. Realistically no one needs more than 80 columns, a prompt and a basic text editor to do real work… or have fun 😁
Seriously now. Can't modern motherboards lock CPU parameters, like frequency, in the BIOS as it is done with RAM *? If I am not mistaken, CPUs are tested in the factory with a fixed (rated) clock speed to see if they are good or not. If those CPUs you bought are tested with fixed parameters, they should all by within a few tenths of percentage from each other.
* Note: I am still on AM3 and I have not had any hands on contact with anything more recent and that is why I am asking. Not even sure if you can still lock RAM parameters. Every video I watch mentions memory training :S
The vegemite you guys get in canadia is very weird looking.
ANOTHER INNACURATE REVIEW?? DID YOU FRESH REINSTALL WINDOWS AND GAMES FOR EVERY CPU?, BECUASE IF YOU DID NOT OF CORUSE THERE WILL BE DIFFERNCES…JUST AS YOU WILL SEE M.2 PEFORMANCE DEGRADING ISSUES.
STOP WITH BULLSHYTE VIDEOS PLEASE
Why not disable the boost/turbo and run them on constant frequency? I'm pretty sure the motherboards still allow this. Of course there will be some performance loss vs maximum possible speed per chip, but if you want consistency, loosing couple percent shouldn't be a big deal. Just manually overclock them to highest frequency the slowest chip can handle without thermal throttling and you should have perfectly consistent result.
I've a solution… don't care so much about a few percentage performance.. If it runs your game smooth, be happy. Instead of researching the best component get a job and upgrade quicker. Enjoy your game, not your hardware (never understood windows in cases either, too bad it returned).
However… I do really appreciate this type of content. Fun to know and have it out in the public, great job!
It's good to see a breakdown of the methodology here considering the previous controversy with LMG's accuracy. I feel as thought this sort of introspective analysis on methodology is showing that they've learnt from their mistakes and are putting in the work.
I love your ambitions! Maybe you could use statistics to ensure one processor series better than the other aka null hypotheses is recected😜
11:48 there is a mistake with the legend : AVG and 1% LOW were inverted
My cpu is now also named Totodile.
4d chesssss
So how about GPU performance variance
This is insanely deep (that's what she said)