We were afraid to publish this video – Intel 13th Gen Review

[Ultraforce]

1 hour ago, AndreiArgeanu said:

The video feels very sloppily made. Like they had to rush in last minute and make the video. I get that the video had to be released around the same time as the others but you could have delayed an hour or two and fixed the mistakes rather than damage your integrity to get some extra views early on. 

Yeah, the Subreddit has mentioned that 5800x3D is listed in the review as having an L3 cache of 25 MB rather than 96 MB of L3 cache.

[kokosnh]

As the graph topic was already mentioned, don't do this type of thing. 
If you avanange the numbers, avarange the graph as well, if not, round it to one decimal point in graph and text...  
This just look's bad. 
it reminds me of these intel, and apple graphs  

[kokosnh]

18 hours ago, Geneleco said:

I really don't like these graphs, it hurts my brain to look at the, need to pause for ages to get a clue of wtf im looking at.

 

gamers nexus are drawing this type of graphs separately, first for one CPU, then for other CPU, while explaining what I'm looking at (it's easier to see)

[dancfuller]

I liked the video! Thanks for including the RAM comparisons.

 

In the future, please add the source file duration to the title of the AV1 encoding graph's title (~8:45 in the video) or change the datapoint shown to average frames per second of the encode. This is a real world datapoint that effectively just turns into a points-based synthetic benchmark if the video duration or fps isn't included. With SVT-AV1 (which Handbrake uses) allowing for reasonable encoding times these days at preset 6 which Handbrake sets as default, knowing how close or far the result is from real-time speed is very useful to know. Another option would be a vertical line on the graph showing where "real-time" lands so there's a very visual "faster than real-time" and "slower than real-time" split. 

 

After that, though there's never a right answer to which CPUs to cross-test, it would've been neat to see the 5950x included in the roundup because the type of person to have bought a 5950x wants to see how their top-of-the-line CPU+platform compares to the latest top-of-the-line CPU+platform. As is, the Ryzen 7000 video needs to be cross-referenced to get that comparison.

 

Thanks!

[thekingofmonks]

Usual LMG cringebait title

[papajo]

On 10/20/2022 at 7:44 PM, James said:

Buy a Ryzen 5800x3d: https://geni.us/A7wA2sX
Buy a Ryzen 7950X: https://geni.us/bSNf
Buy a Ryzen 7600X: https://geni.us/bZGspc
Buy a Ryzen 7900X: https://geni.us/zZBOzJ
Buy an Intel 12600K: https://geni.us/TJQqCTC
Buy an Intel 13600K: https://howl.me/cikwLTUaN9I
Buy an Intel 13900K: https://howl.me/cikwLVdik3b

Purchases made through some store links may provide some compensation to Linus Media Group.

 

AMD ushered in the new generation of desktop CPUs just last month, but Intel is clapping back in a major way. Can Team Red afford to sit on their Ryzen laurels? Or will 13th gen give them a run for their money… and customers?

 

 

 

That's some BS talk right there 

-To "help" each platform put its best foot forward we paired them with the "best" kits of ram as recommended by each manufacturer.... 







LTT always tries to hinder AMD and boost intel every single time with stupid details like that... 

1) Everybody and his grandmah knows that faster ram = better (although marginal in most cases) performance in games 
2) Everybody and his grandmah knows that best ram for ryzen 5 (best in the sense that it will be 99% compatible and run without issues no matter the mobo/cpu combo) is DDR4 3800Mhz 

3) We already know that ryzen 7 has performance increases in games if you pair it with faster ram here is a hardware unboxed comparison for the 7600x using a variety of ram speeds starting from DDR5 5600 and ending on DDR5 6400  also notice that the differences in the ltt video  of this topic could be nullified if ryzen had the performance difference we see in the hardware unboxed using a 6400 Mhz ram and probably would have even better results if it was to use a 6800mhz ram, 

Also notice that the same happened in the hardware unboxed videos where when they used the same speed of ram (6400 for both intel and 7600 ) instead of 6400 for intel and 6000 for amd most benches where intel came ahead were reversed. 


Last but not least this is for the 7600 I would expect higher clocked CPUs (7900, 7950) to have an even better boost if paired with faster memory.  (simply because* FPS means how much latency the CPU has on each single frame if the latency drops with higher ram speed it will drop further with higher clock speed as well!

*and assume that GPU latency is way lower to not play a factor so in other words in a CPU heavy 1080p benchmark with a moster GPU like a 4090) 
 

4) you did not tell us what fabric to ram ratio you used was the ram dual rank or single rank and the setup in general  e.g how was the PBO configured?  was the cooling system the same for both CPUs for example? and which model did you use for a CPU cooler ?

I bet even there you used a better cooler for intel because you do stuff like that at ltt either out of negligence or because of providence 

BTW the difference in money from the cheapest Gskill 6000mhz to the cheapest gskill 6800mhz is about 170 usd  and they say intel performs better (4% while having faster ram and with all the other "ltt bottlenecks " anti amd benchmark tricks of ltt (e.g maybe they disabled PBO because it is supposedly OC... but forgot to turn off any ASUS "default" bios setting boosts on the intel side) ) while having a 150 usd price gap between the CPUs....  why would one say that if in order to get that result you need to pay +170 euro for the ram either way lol ....

[papajo]

On 10/20/2022 at 7:51 PM, emosun said:

At least it's not as bad as the 90's where you'd literally see your flagship cpu become half the speed of the next flagship only a year later.

Yea those were very bad times each new year CPUs were noticeably faster and also CHEAPER!! 


Now it is better where speed differences (especially if AMD wasnt to introduce the ryzen line) are negligible or small and the prices double each year... yea NOW are the good times  
 

  

On 10/21/2022 at 4:33 PM, Geneleco said:

I really don't like these graphs, it hurts my brain to look at the, need to pause for ages to get a clue of wtf im looking at.

It is easy dots are for how much power each CPU used  at a given time in seconds (while running prime 95) 

Lines are at what temperature that CPU was at a given time in seconds (while running prime 95) 

In other words dots use the bottom(time) and right(power) labels of the graph

lines use the bottom (time) and left (temp) labels of the graph 

both dots and lines have a common variable (time) 

And a rule of thump for these graphs if there are labels both left AND right

then there must be two different symbols (e.g here dots and lines) of the SAME color (probably referring to the same product) and one symbol references the right label (here the dots reference the right label which is power) and the other symbol references the left label (here the lines reference the label on the left which is temperature) 

and all the symbols will have a common variable on the bottom (in our case it means that they have a particular power value and temp value on a particular second, you find that value by choosing the second you like e.g 421 and draw a straight vertical line and see which dot and line it hit and then see left for temp and right for power values ) 

Having said that I will agree that it would be easier to the eye if they had separated the symbols and colors by grouping them per product instead of per value type.. 

so just switching up the lables on the top side like this: 



And I just noticed that there is a labelling error it should have "12600K" for the blue dots instead of "13600K" lol 



So like this: 



Or even maybe something like this: 

[emosun]

1 hour ago, papajo said:

Yea those were very bad times each new year CPUs were noticeably faster and also CHEAPER!! 


Now it is better where speed differences (especially if AMD wasnt to introduce the ryzen line) are negligible or small and the prices double each year... yea NOW are the good times  

 

[papajo]

3 minutes ago, emosun said:

 

Your point 



Me 

[MultiGamerClub]

On 10/20/2022 at 6:51 PM, emosun said:

At least it's not as bad as the 90's where you'd literally see your flagship cpu become half the speed of the next flagship only a year later.

Reminds me in 2003 when i got my first HP laptop, fastest on the block.. for 2 weeks i think.. Good times.

[emosun]

1 hour ago, papajo said:

Your point 

I think what i like most about this reply is knowing you either went back and checked and still didn't figure it out , or you didn't actually check a second time. Either way it's not looking to great for you atm but i'd highly recommend maybe taking another 18 days past the thread being posted to formulate a response.

[emosun]

56 minutes ago, MultiGamerClub said:

Reminds me in 2003 when i got my first HP laptop, fastest on the block.. for 2 weeks i think.. Good times.

yeah. the worst I saw was in in 2 years going from 450mhz to 1.5ghz from 1998 to 2000. just absolutely killing off anyone who bought something high end only 2 years prior

[Sarra]

2 hours ago, emosun said:

yeah. the worst I saw was in in 2 years going from 450mhz to 1.5ghz from 1998 to 2000. just absolutely killing off anyone who bought something high end only 2 years prior

Yeah, those years where performance went up by like 5% over a 5 year period were pretty boring.

 

I mean, competition really stifles the market and all.

[Shimejii]

5 hours ago, papajo said:

2) Everybody and his grandmah knows that best ram for ryzen 5 (best in the sense that it will be 99% compatible and run without issues no matter the mobo/cpu combo) is DDR4 3800Mhz 

This is Just wrong. Plain and simple wrong. Most Ryzen systems from 3000 and 5000 series pretty much only support up to 3600, beyond that its not guaranteed, especially not that 99% compatibility crap. Hell most systems ive helped build for most people their CPUS didnt even like the QVL Ram at 3600 mhz quite often and they had to be turned down to 3200mhz to be the most stable, shit happens.

 

Intels 12th Gen IMC was pretty awful and had a ton of issues going beyond 6000, or doing pretty much any 4x16 or 4x32 kit of ram. 13th Gen fixed a ton of that and they should be able to use MUCH faster memory than AMD. AMD has always been step or two behind Intel in terms of RAM compatibility and speed with ryzen, 1st and 2nd gen had some pretty awful IMC's, 3rd and 5th gen were much better but even they had their limitations.

 

Blanket statements like this just make your entire argument just look wrong. You have valid points, as ive stated unless you test them all using the same ram kits (in this case it would have been 6000) then these kind of tests are pretty much lose a fair amount of value. LTT has not been known for their accurate testing methods, or the go to for accurate benchmarking for years now, and labs is not gonna fix that stigma with the way they are trying to automate it. But its just another data point to look at, and more information in your hands gives you a better opportunity to make a better educated decision. 

 

In the end Ryzen 7000 is a great leap forward, but the pricing of the lower two chips is shit, there is no getting around that. Their boards are 100$ overpriced in most spots, and they kind of just didnt take that much of a leap compared to their 5800x3D. Thats just shows you that the 5800x3D is just a quality CPU for gaming, and its good enough for most people for a while until the upgrades matter enough and it comes time to upgrade. Usually this is every 5 years or so for most, some a bit sooner some later.

 

13th gen is a very good lineup and is a lot more price competitive this time around, there is no getting around that. But its can also be the power hog, and both get pretty hot but 13th gen can thermal throttle quite easily if not properly done. Each has pro's and cons.

[MultiGamerClub]

5 hours ago, emosun said:

yeah. the worst I saw was in in 2 years going from 450mhz to 1.5ghz from 1998 to 2000. just absolutely killing off anyone who bought something high end only 2 years prior

Saw the same shit when i was an amd fanboy back in the early days.. had a Radeon 7700HD and it ran good, then nvidia just drops an f-bomb and nuke the performance no sweat.. Felt kinda like shit.

[papajo]

4 hours ago, Shimejii said:

This is Just wrong. Plain and simple wrong. Most Ryzen systems from 3000 and 5000 series pretty much only support up to 3600, beyond that its not guaranteed,

Putting ryzen 3000 and 5000 in one group tells me that you dont know what you are talking about. 

all 5000 can do 3800 1:1 

And the 5000G can do even higher. 

[IkeaGnome]

36 minutes ago, papajo said:

Putting ryzen 3000 and 5000 in one group tells me that you dont know what you are talking about. 

all 5000 can do 3800 1:1 

And the 5000G can do even higher. 

You're missing this. They went with Intel and AMD's listed speed ratings. If they didn't then it's "They overclocked ram for AMD, but not Intel" or "They overclocked the ram more for AMD than Intel" It's not reviewers' fault that AMD has a lower listed speed on their CPUs anymore.

According to AMD, no the 5600g doesn't go higher. 

AMD lists 5200 mhz for the 7600x. That's what it was tested at right?

Want to guess what speed they tested 13700k at?

[Shimejii]

1 hour ago, papajo said:

Putting ryzen 3000 and 5000 in one group tells me that you dont know what you are talking about. 

all 5000 can do 3800 1:1 

And the 5000G can do even higher. 

Tell me you dont build systems without telling me you dont build them. You just blankly read spec sheets  No its not every single one, its a good amount of them, but not every one. Is it likely they will get 3600? Yes with the right memory kit and board it should, but its not guaranteed. Blanket statements like that would easily get you in trouble if you had any sort of reputable business. 

 

The IMC itself was nearly identical between 3000 and 5000 series, they did tweak it a slight bit and that helped it get to be that stable, but they are pretty much the same IMC. With a Year and a bit of making them they tuned it a bit better for 5000 series, but it still had pretty similar issues. 

 

Again out of a few dozen systems ive built on both 3000 and 5000, about 90% of them got 3600 mhz no problem, 3800? No only about half of them passed that with just the XMP enabled with a 3800mhz kit. Customers dont want to hear about how you have to fine tune it in the bios if anything happens, they just want it to work. If they want to do it, more then likely they would have just built their own system. At that time i had tried 3 kits out, gskill, Corsair and Crucial, and i had a lot  more luck with crucial working on most of the ones that worked. 

[Geneleco]

17 hours ago, papajo said:

It is easy dots are for how much power each CPU used  at a given time in seconds (while running prime 95) 

Lines are at what temperature that CPU was at a given time in seconds (while running prime 95) 

In other words dots use the bottom(time) and right(power) labels of the graph

lines use the bottom (time) and left (temp) labels of the graph 

both dots and lines have a common variable (time) 

And a rule of thump for these graphs if there are labels both left AND right

then there must be two different symbols (e.g here dots and lines) of the SAME color (probably referring to the same product) and one symbol references the right label (here the dots reference the right label which is power) and the other symbol references the left label (here the lines reference the label on the left which is temperature) 

and all the symbols will have a common variable on the bottom (in our case it means that they have a particular power value and temp value on a particular second, you find that value by choosing the second you like e.g 421 and draw a straight vertical line and see which dot and line it hit and then see left for temp and right for power

Yes thank you i know how to read the graph, its just harder than it needs to be. The labeling might help, but its def a combination of the symbols used for plotting, the background colour and if we even really want to plot both those things on one graph instead of 2

[papajo]

On 11/9/2022 at 11:06 AM, Shimejii said:

Tell me you dont build systems without telling me you dont build them.

 

On 11/9/2022 at 5:22 AM, Shimejii said:

Most Ryzen systems from 3000 and 5000 series pretty much only support up to 3600

 

On 11/9/2022 at 5:22 AM, Shimejii said:

Hell most systems ive helped build for most people their CPUS didnt even like the QVL Ram at 3600 mhz quite often and they had to be turned down to 3200mhz to be the most stable, shit happens.

 

Putting stuff together (that btw is made so that even children can build them ) doesnt make you knowledgeable <---- that's for even if your claims about building PCs is true. 

The two generations have different PHYs so even if your ridiculous claim about 3600mhz not working (obviously on 3000) is of your user error (in case it was indeed QVL) or your incompetence (in case it was not QVL) not being able to figure out how to set secondary and/or tertiary timings etc. 
 

Riddle me this, every time someone comes in your "shop" to fix his computer because it is slow or has viruses or whatever what you do is format the disk and reinstall windows right? lol 

[powertoys]

21 hours ago, papajo said:

Riddle me this, every time someone comes in your "shop" to fix his computer because it is slow or has viruses or whatever what you do is format the disk and reinstall windows right? lol 

 

I'm not the person you're replying to, but a clean install of Windows is by far the best (and fastest) solution to get rid of every single trace of malware in the OS... of course, if the customer doesn't mind a factory reset.

 

And nearly all complaints I got about "computer too slow" was because the computer manufacturer was doing a disservice to themselves by selling machines with Windows 10/11 + 5400rpm HDD. Upgrading to SSD fixed every single one of these computers, and bonus points if the customer allows you to clean install Windows in the SSD (instead of cloning the existing OS) to get rid of all OEM bloatware.

[papajo]

15 hours ago, unnerfed cthun said:

m not the person you're replying to, but a clean install of Windows is by far the best (and fastest) solution to get rid of every single trace of malware in the OS

Yea I was referring to the guy that thinks ryzen 3000 and ryzen 5000 series are the same in terms of ram compatibility/settings 

As for formatting a computer being the best solution, It is not because people want to keep their settings files and programs + many times malware is not the reason for issues, in fact best practice is to warn your customer if you gonna have to format his computer. 

Not everyone uses his computer just for facebook and gaming, there is custom software (especially for businesses) which license can not be replaced -or doesnt make financial sense to replace) for various reasons -company doesnt exist anymore or newer version is out that replaces the old one and doesnt have compatible license with the old one so one has to buy the new one or the guy who coded it doesnt care after e.g 5 years to answer an email or whatnot and reinstall his software and so on and so forth) and will get lost if the program is removed/uninstalled.


Now if it is YOUR(not you personally per se, you know what I mean) computer and YOU are going to fix it I can understand that formating a PC is the most convenient solution many times especially the less things you know about what happened to your computer/how to monitor and check it or ways to reverse it.