Regarding the recent Tech Quickie

[Prysin]

@MageTank its happening again...

Dunno if you noticed, but @LinusTech made yet ANOTHER video spreading partially false information about the effect faster/better timed RAM has on even normal day to day tasks.

 

In his video (availible on Vessel) link below:

https://www.vessel.com/videos/f3JMXRNiP?tab=details

 

He yet again claims RAM has near no impact on gaming performance. Whilst on average the boost may not be big, massively popular titles such as, but not limited to; Fallout 4, Ryse Son of Rome, GTA V and The Witcher 3 will see noticeable performance increases by going with faster ram, ESPECIALLY, on lower or mid end CPUs (Athlon 860k, FX 4/6/8/9 series, Core i3s, locked i5s, quad core no HT Xeons). Why is this? Because said CPUs have downsides to their performance, and any sort of decrease in access time will result in directly measurable performance increase.

 

There has been several documented cases of how and when RAM speed and latency improvements offers benefits. By representing his video the way he does, Linus make it sound like the 5-15 USD more you pay for a kit of faster RAM is more like 50-150 USD more, by constantly referring to faster RAM as a "much more expensive choice".

 

The fact that @LinusTech goes as far as to promote estetic's over performance, WHILST claiming that the "enthusiast" segment would agree that spending the money on higher speed is wasted is not just wrong. It is simply sickening. Anything that can increase performance without derailing a budget should be considered as a serious option. Picking slower RAM with LEDs on them offer no benefit at all. It looks ballin, but performs the same OR LESS due to generally worse specifications on kits with LEDs or other fancy effects.

 

At the same time, Linus's builds sometimes feature faster RAM kits, which is often given as a sponsorship or has been bought at some point. But whenever faster RAM is mentioned, it is always brushed aside as if it has no performance benefit.

 

So @LinusTech, Tell me. At what grounds can you claim that the performance increase you get with faster memory is not worth it after reading / watching the following documentation

 

 

 

 

 

 

http://www.overclock.net/t/1487162/an-independent-study-does-the-speed-of-ram-directly-affect-fps-during-high-cpu-overhead-scenarios

 

 

 

 

@Slick@nicklmg (tagging you guys too so this maybe gets read.)

I also have one more grief with the video, that sounds dumb but to the unknowing computer user, it may not be obvious.

When Linus was talking about AMD APUs using system memory and dGPUs (read: graphics cards) using their own memory, a slide of a Nvidia dGPU was shown. This CAN cause the misconception that AMD uses system memory which is slow and terrible and bad (blablabla insert fanboy arguments here).

There should have been a brand-agnostic example slide not showing ANY logo, serial code or branding. OR both brands should have been equally represented with a slide each, as to not cause potential confusion and or misconceptions.

Why can this cause misconceptions or confusion?

Because @LinusTech never really explained what an APU is, he never clarified if it was a graphics card or a CPU with iGPU. Hell, even AMD doesnt clearly call it an APU. Nor does he explain what APUs are commonly used for or where you commonly find them.

AMD APU products are labeled A6, A8 or A10 followed by four numbers representing the generation and tier within the product stack. By not being clear on this point, Linus is not being clear to his less knowledgable audience.

I hope LMG edits said slide to remove the Nvidia branding before this goes onto "mainstream" platforms such as Youtube as to avoid any ignorant idiots using it as an real argument for or against a product or brand in a open discussion.

[Enderman]

Is 2 fps really that big of an improvement?

I mean if you get higher speed ram at the same price then sure, but if youre spending $100+ to get higher speed ram you would get more performance by spending that on a better GPU or CPU

[pagani123]

It is generally considered among the community to not be a big enough improvement in gaming to count it as worth spending more money on. Also the thing about the apu and showing nvidia dgpu is just nitpicking and very fan-boyish.

Edit: Also the sources your citing use a few generations behind micro architecture which means in the newer architecture you might not see as high of performance gain as they are reporting.

[don_svetlio]

8 minutes ago, Maybach123 said:

It is generally considered among the community to not be a big enough improvement in gaming to count it as worth spending more money on. Also the thing about the apu and showing nvidia vram is just nitpicking and very fan-boyish 

Trust me, there are dumb people who will claim what Prysin predicted

[Enderman]

10 minutes ago, ONOTech said:

Where do you see 2 FPS? All of the links OP posted show a pretty noticeable jump. Like 10+

You're looking at the benchmarks wrong.

You need to compare between the SAME CPU

there are two different i3s there not one

if you compare the same CPU there is pretty much always a 2fps difference

[pagani123]

I assume that means your siding with me correct? (sorry pulled an all-nighter studying and i might not be reading things straight)

[don_svetlio]

3 minutes ago, Maybach123 said:

I assume that means your siding with me correct? (sorry pulled an all-nighter studying and i might not be reading things straight)

I agree with Prysin - when comparing APU VRAM and GPU VRAM no brands should be mentioned

[pagani123]

Just now, don_svetlio said:

I agree with Prysin - when comparing APU VRAM and GPU VRAM no brands should be mentioned

ahh now i understand you were siding with him on the part that i said was a bit nit picky 

[Enderman]

7 minutes ago, ONOTech said:

Umm..?

 

6-10 FPS, which is still pretty minimal under certain circumstances, but it's definitely not as small as you claim.

i was watching the second video in the OP and the difference is almost always 2fps

[don_svetlio]

3 minutes ago, Maybach123 said:

ahh now i understand you were siding with him on the part that i said was a bit nit picky 

He is right though. This particular subject has no connection with GPU brands and they should not be mentioned so as to not confuse the reader/viewer with pointless information

[AshleyAshes]

I'm mostly just enjoying that last video where the i5 2500K is shown to more or less still hold it's own five years later.  Sure they point out some weaknesses but those weaknesses are pretty minimal if you consider that this is a CPU that cost only $200 five freakin' years ago.  To get 'pretty good performance' out of a CPU five years down the line is a massive return on your investment in a piece of tech.

[Prysin]

1 hour ago, Enderman said:

Is 2 fps really that big of an improvement?

I mean if you get higher speed ram at the same price then sure, but if youre spending $100+ to get higher speed ram you would get more performance by spending that on a better GPU or CPU

going from 1600MHz DDR3 2x4GB kit to 2400MHz DDR3 2x4GB kit costs 7 USD.

Going from 2133MHz DDR4 2x4GB kit to 2666/3000MHz DDR4 2x4GB kit costs 13-15 USD.

 

Digital foundry has clear evidence of the performance increase.

 

Similar performance increase (5-10 FPS) if we were to use other components to increase performance.

If upgrading CPU to get 5-10 FPS (assuming infinetively powerful GPU and bog standard RAM) -> 50-100 USD

if upgrading GPU to get 5-10 FPS (assuming infinetively powerful CPU and bog standard RAM) -> 50-100 USD

 

If the system is using a lower end CPU, like what i mentioned in the OP, the cost of getting 5-10 FPS without overclocking (if you can) is equal to the cost of upgrading to a Core i5 2500k, 3570k, 4690k or 6600k and then overclock said CPUs from whatever you are currently using. Which means you must invest in a better cooler (212 Evo = 30 USD), potentially better mobo (assuming you sell your current mobo at break even) +20-30 USD....

 

Aslong as a mobo supports it, there is no good argument against faster RAM other then "i dont wanna spend 15 bucks more".....

[Prysin]

2 hours ago, Maybach123 said:

It is generally considered among the community to not be a big enough improvement in gaming to count it as worth spending more money on. Also the thing about the apu and showing nvidia dgpu is just nitpicking and very fan-boyish.

Edit: Also the sources your citing use a few generations behind micro architecture which means in the newer architecture you might not see as high of performance gain as they are reporting.

The sources i am using is showing Sandy Bridge (Video #3), Haswell (video #2), AMD FX (Piledriver) (Video #2) and Skylake (Video #1 and #2). I have every currently relevant and or in use architecture except Westmere/Nehalem represented here.

 

Ivy Bridge is represented in the OCN thread by a 3930k Ivy Bridge-E system.

 

Unless the video is sponsored by say Nvidia and it is explicitly stated, their products unless specifically relevant to the content of the video should not be used as an example as it CAN cause confusion, or retarded fanboys will twist it into a moronically and ignorant argument that only a moronic retarded fanboy could even come up with.

having a Nvidia branded dGPU sample had NO relevance to the content of the video other then to feed fanboys munition to create retarded arguments.

 

Linus Media Group has on previous occasions used brand agnostic models (3D renders/PCBs without branding) for their Techquickie videos. Why they did not use it this time, i do not know. But it was the wrong move. They have the samples needed to NOT use potentially misleading branding.

[MageTank]

I currently do not have a Vessel account, but I feel like I am going to need to have a sitdown with @LinusTech on the subject of memory. While I have not seen this video, I am curious if he has actually taken the time to test the impact of speed and specific timings on games (Like I have) and is willing to even consider putting aside that time to do a proper test. Not just load a random XMP profile on a few random titles, use a bottlenecked GPU, and record average framerates and conclude that ram speed is worthless.

 

I am working on a very large (several page long) comprehensive guide to memory overclocking and its impact on not just gaming, but hardware in general (Including SLI and Crossfire setups and the impact memory latency has on these two things). I am waiting on more DX12 titles to test, but if need be, I can release what I have now in an attempt to sway people back into believing memory speed is important. Seeing as the price difference between a 2133 and 2666mhz kit is only $4 on 16GB kits, there really is no reason to use 2133mhz. @Prysin can you tell me if he went over tertiary timings and RTL/IO-L in that Fast As Possible video? 

[Prysin]

6 minutes ago, MageTank said:

I currently do not have a Vessel account, but I feel like I am going to need to have a sitdown with @LinusTech on the subject of memory. While I have not seen this video, I am curious if he has actually taken the time to test the impact of speed and specific timings on games (Like I have) and is willing to even consider putting aside that time to do a proper test. Not just load a random XMP profile on a few random titles, use a bottlenecked GPU, and record average framerates and conclude that ram speed is worthless.

 

I am working on a very large (several page long) comprehensive guide to memory overclocking and its impact on not just gaming, but hardware in general (Including SLI and Crossfire setups and the impact memory latency has on these two things). I am waiting on more DX12 titles to test, but if need be, I can release what I have now in an attempt to sway people back into believing memory speed is important. Seeing as the price difference between a 2133 and 2666mhz kit is only $4 on 16GB kits, there really is no reason to use 2133mhz. @Prysin can you tell me if he went over tertiary timings and RTL/IO-L in that Fast As Possible video? 

he merely mentioned latency. The closest he got was to mention that "some types of RAM (Random Access Memory) has a string of numbers representing the latency like 9-9-9-24".

 

He didnt even get any deeper into explaining anything.

 

To be honest, given how complicated memory timings can be to learn (assuming one know not a damn shit about computers. Which is the kind of audience Techquickie is aiming for) the video wasnt even worthy of being called a "overview".

With his old memory speed video in mind, it was more like a "second serving" of false or half arsed info.

[MageTank]

1 minute ago, Prysin said:

he merely mentioned latency. The closest he got was to mention that "some types of RAM (Random Access Memory) has a string of numbers representing the latency like 9-9-9-24".

 

He didnt even get any deeper into explaining anything.

 

To be honest, given how complicated memory timings can be to learn (assuming one know not a damn shit about computers. Which is the kind of audience Techquickie is aiming for) the video wasnt even worthy of being called a "overview".

With his old memory speed video in mind, it was more like a "second serving" of false or half arsed info.

Then I guess my guide will have to come early. I'll also have to make a video with it, as kids these days are too good for reading things. Linus might have the charisma to win people over, but even that is no match for a seductive MageTank voice.

[Prysin]

Just now, MageTank said:

Then I guess my guide will have to come early. I'll also have to make a video with it, as kids these days are too good for reading things. Linus might have the charisma to win people over, but even that is no match for a seductive MageTank voice.

what kinda CF setups do you have availible?

I got my R9 295x2

also got a 950 which i use alongside a Athlon X4 845 (desktop carrizo)... however i have to manually OC the 845, as XMP OCs the locked CPU causing the BIOS to reject the settings upon reboot.

[MageTank]

4 minutes ago, Prysin said:

what kinda CF setups do you have availible?

I got my R9 295x2

also got a 950 which i use alongside a Athlon X4 845 (desktop carrizo)... however i have to manually OC the 845, as XMP OCs the locked CPU causing the BIOS to reject the settings upon reboot.

Currently don't have any. A friend of mine was gonna handle that part of the test, but I can invest in a couple AMD cards if need be for the test. I have Newegg Premier which means I have no restocking fee (kind of a terrible thing to do, but it gets the job done). I do have laptops with Hybrid Crossfire, that I can try to test with. 

 

For SLI, the tests are done with 1, 2, and 3 GTX 780's (Testing single card, dual SLI, and triple SLI) and we have a few extra cards to test with across multiple generations (GTX 560 Ti, GTX 750 Ti, GTX 770, GTX 970). Trying to show if budget machines are better off investing in faster memory or not, when GPU is bound to become the bottleneck much faster. Such information is extremely helpful, so i hope to include as much details as physically possible (Along with teaching people absolutely everything I know about each individual ram timing, what timings impact each other, and their impact on bandwidth and latency in general). Once this guide is finished, an amateur should feel confident enough to manually overclock their ram assuming they have enough time set aside to do so.

 

After that, the "faster ram is too expensive" argument will be completely nullified.

[Citadelen]

4 hours ago, Enderman said:

-snip-

Do you always speak before you think?

[Prysin]

1 minute ago, Citadelen said:

Do you always speak before you think?

his mind may be absent, but his fingers are not.

[Prysin]

8 minutes ago, MageTank said:

Currently don't have any. A friend of mine was gonna handle that part of the test, but I can invest in a couple AMD cards if need be for the test. I have Newegg Premier which means I have no restocking fee (kind of a terrible thing to do, but it gets the job done). I do have laptops with Hybrid Crossfire, that I can try to test with. 

 

For SLI, the tests are done with 1, 2, and 3 GTX 780's (Testing single card, dual SLI, and triple SLI) and we have a few extra cards to test with across multiple generations (GTX 560 Ti, GTX 750 Ti, GTX 770, GTX 970). Trying to show if budget machines are better off investing in faster memory or not, when GPU is bound to become the bottleneck much faster. Such information is extremely helpful, so i hope to include as much details as physically possible (Along with teaching people absolutely everything I know about each individual ram timing, what timings impact each other, and their impact on bandwidth and latency in general). Once this guide is finished, an amateur should feel confident enough to manually overclock their ram assuming they have enough time set aside to do so.

 

After that, the "faster ram is too expensive" argument will be completely nullified.

That argument will forever be there.

Just like the argument that "AMDs drivers are horrible", because back in 2006-2010 they were indeed really horrible.

[Enderman]

1 hour ago, Citadelen said:

Do you always speak before you think?

Did you not watch the videos OP posted before making rude comments?

1 hour ago, Prysin said:

his mind may be absent, but his fingers are not.

I said that spending lots of money on higher speed ram is not worth it

I also said that if its not much more expensive, then it is

So if you think im wrong, does that mean that you believe spending tons of money on faster ram for 2-4fps is worth it?

Are your eyes absent? Or are you just trying to be an asshole because that's what your parents taught you to do?

 

[Suika]

5 hours ago, Prysin said:

...noticeable performance increases by going with faster ram, ESPECIALLY, on lower or mid end CPUs (Athlon 860k, FX 4/6/8/9 series, Core i3s, locked i5s, quad core no HT Xeons)

 

...

 

WHILST claiming that the "enthusiast" segment would agree that spending the money on higher speed is wasted is not just wrong.

To be fair, the enthusiast segment typically doesn't use lower or middle end processors, so in many situations a hardware enthusiast wouldn't see the difference.

[Citadelen]

7 minutes ago, Enderman said:

-snip-

Hey you've made comments about me before, I wanted some fun too. Also, hypocrisy much, 'Or are you just trying to be an asshole because that's what your parents taught you to do?'

[Enderman]

1 minute ago, Citadelen said:

Hey you've made comments about me before, I wanted some fun too. Also, hypocrisy much, 'Or are you just trying to be an asshole because that's what your parents taught you to do?'

What comments?

All I said was that faster ram isnt worth it if it costs a lot more, but i its only a little more than it is.

Like, what's wrong with what I said? Are you upset for no reason? Having a bad day?