Re: Let's talk about our big mistake... - AMD Ryzen 2nd vs 3rd Gen - LTT, can you stop flip flopping please?

[Drak3]

3 minutes ago, BlueGoliath said:

Maintain at all times?

My 5930K has no issues maintaining 80 at stock with a 1080, and that is in the same ballpark as the R5 1600X.

1 minute ago, Arika S said:

I feel like this was supposed to be some sort of "gotcha" moment, but i have no idea what you're trying to get at.

 

It’s a poorly executed “You believe one thing (that I believe) is bad and wrongthink, therefore everything you say is bad and wrongthink.”

 

[BlueGoliath]

3 minutes ago, 2Buck said:

I'd really investigate what's causing that, because that's pretty messed up. I'm pretty sure if I lowered the settings enough my 3.4ghz q6600 rig could hack that. Not saying that as an absolute fact, but I'm pretty damn sure.

 

I already know what's causing it: memory latency. You look at MSI afterburner in the city and no single CPU core is being utilized 100% while the GPU is sitting there at like 30-40% and will even downclock due to lack of load. In the desert the opposite is true and FPS is more reasonable. You fly towards the city while looking at the city and the FPS plumments to like 45. Benchmarks look better because they are in the desert mostly, not the city.

 

You don't see this with an Intel CPU because it doesn't have 80ns of memory latency.

 

I tried tightening the timing and disabling SMT to reduce latency as much as possible(61ns at one point according to AIDA) which got me massive FPS gains but as Buildzoid said, getting memory overclockings stable on the Taichi is a pain in the ass even if it POSTs. Well, that and I don't have a clue how to do subtimings.

[GoldenLag]

18 hours ago, BlueGoliath said:

mean, the 1800x isn't a good gaming CPU by a long shot but it's very good at software compiling which I do more often than you'd think

Very few people liked the 1800x. Especially since the 1700 was a thing....

 

18 hours ago, BlueGoliath said:

Get? Yes. Maintain at all times? No.

So its a frametime issue?

 

Then its still quite possibly a setup issue. 

 

18 hours ago, BlueGoliath said:

 

I already know what's causing it: memory latency

What frequency are you running?

Is it singlechannel?

You mentioned twealing timing, what did you achieve? 

[Guest]

2 hours ago, BlueGoliath said:

 

I already know what's causing it: memory latency. You look at MSI afterburner in the city and no single CPU core is being utilized 100% while the GPU is sitting there at like 30-40% and will even downclock due to lack of load. In the desert the opposite is true and FPS is more reasonable. You fly towards the city while looking at the city and the FPS plumments to like 45. Benchmarks look better because they are in the desert mostly, not the city.

Not too sure about memory latency when 100% of the GPU is being utilised. Could be a bottleneck I reck in. 

[TheDelphiDude]

4 hours ago, DrMacintosh said:

AMD is simply the CPU of choice for most people in 2019.

Perhaps for enthusiasts, but for the vast majority of PC/laptop owners? They are either unaware of the various brands and manufacturers on offer or simply don't care.

[Zando_]

oof can't read the whole massive reeee text wall, but I can tell you the A320 thing is probably due to that being a low end chipset. And being a low end chipset, OEMs skimped on every point possible from the VRMs to the board itself and the BIOS and everything. Not worth giving A320 Zen 2 support when it'll just kablooie as soon as you put a decent CPU in there. 

Also they promised support for AM4. Intel has been running LGA1151 for a while now and they've gone through multiple non-compatible chipset jumps with all their CPU generations. AMD is still supporting the socket like they said they would, but dropping support for one of the shittiest chipsets ever isn't really an issue and isn't breaking any promises I saw them make. 

[BlueGoliath]

@Crunchy Dragon the "flaming" is caused by trolling of other forum members. Why punish me by locking the thread instead of removing their troll comments and giving them a warning?  You still have people attempting to derail it by suggesting there is something wrong with my setup, which there is not nor is it the point of this topic, yet i'm going to be punished for it?

 

Ryzen first generation was never "top-tier performance". The statements made in the newest video were a blatant lie and there is video evidence to prove it, which is kinda the topic here among other things.

[Crunchy Dragon]

4 minutes ago, BlueGoliath said:

@Crunchy Dragon the "flaming" is caused by trolling of other forum members. Why punish me by locking the thread instead of removing their troll comments and giving them a warning?  You still have people attempting to derail it by suggesting there is something wrong with my setup, which there is not nor is it the point of this topic, yet i'm going to be punished for it?

 

Ryzen first generation was never "top-tier performance". The statements made in the newest video were a blatant lie and there is video evidence to prove it, which is kinda the topic here among other things.

I'm not going to discuss this publicly. If you wish to have more details, feel free to send me a message and we can have a civil discussion about it.

[Sauron]

11 hours ago, DrMacintosh said:

That's funny because I never claimed that it did. LTT already explained why they went with AMD for their LAN center. Ryzen simply brings more to the table and for less money. It's nothing personal dude. You don't have to get so defensive about what CPUs LTT uses. It's not your money and it's not your rig. 

LTT: *uses Intel on a build*

People like this dude: "Shills! AMD eats Intel alive, we all know you're getting money from team blue!

LTT: *uses AMD instead*

This dude: "Shills! Intel outperforms AMD in 2 games, stop taking Lisa's dirty money!"

10 hours ago, BlueGoliath said:

Get? Yes. Maintain at all times? No.

 

 

 

If you're going to complain about it dipping to 58 once every 5 hours you don't have your priorities in order. Oh, and you backed off from what was a straight up lie:

10 hours ago, BlueGoliath said:

the 1800x can't even achieve stable 60FPS in GTA 5 online

yes, the 1800x can ABSOLUTELY achieve stable 60 in gta v.

10 hours ago, BlueGoliath said:

This isn't about reviews but coverage of it in general. Point me to a tech reviewer that mentions BIOS/firmware bugs, please. 

Level1Techs, der8auer, I'm sure I could find more if I watched more. LTT also mentions firmware and BIOS problems sometimes, I'm pretty sure they talked about the problems with x570 multiple times when explaining why they didn't like weekend releases.

[Drak3]

20 minutes ago, BlueGoliath said:

Ryzen first generation was never "top-tier performance".

Yes, it was. It just wasn't the performance king.

24 minutes ago, BlueGoliath said:

You still have people attempting to derail it by suggesting there is something wrong with my setup,

Because your using "your setup" as be all end all to your argument to the 1800X, when it's extremely easy to disprove.

[minibois]

Disclaimer: I consider myself an AMD fan. Not a fanboy, a fan. Despite my preference, I can still be realistic about this world of tech.

 

 

A320 is a bullshit chipset, which should have never been made. B350 was only a bit more money, even at launch of the chipset and to me it was clear from the start A320 was not going to be fit for later generations of products. It surprised me at the time even AMD let A320 support CPU's above the base quad and hexa cores. Not even mentioning it supporting Ryzen 7 8-core CPU's...

I am not surprised that chipset didn't get an upgrade to Zen 2, but I think it is a fair criticism towards AMD. They had said AM4 will last until 2020 (if memory serves me right), but no asterisks behind that statement.

 

AMD did do a lot of things quite right, like pressuring Intel to give people on a consumer platform more cores. They did this by releasing CPU's that double the threads or cores in a given price point, while not matching single core performance.

That meant their performance was not the best on gaming, but it would depend on the system you're running. For example a 1080p 144Hz system would struggle more on Ryzen, than a 4K 60Hz system would.

 

Still though, gaming is not all and with certain workloads that leverage many CPU cores (e.g. video rendering, streaming/recording, etc.) AMD was a really good choice. Even if Intel had an answer on their HEDT platform, AMD had a lot more value.

You could make the argument in many cases that GPU rendering (e.g. in Adobe Premiere or NVENC streaming/recording) was an answer to that.

 

I don't think it's a lie that AMD had great performance over the past generations, especially in Zen 2/Ryzen 3000. In the past that single core performance was kind of lackluster, but the added core amounts did make it a viable system choice in many regards.

 

 

Although I can from experience say certain motherboards were kind of ... Difficult to work with, certainly in regards to RAM speed compatibility.. Which I can surely see as a turn off to many people.

 

On topic on the main statement (that about claiming 'top tier performance'): the difference in performance between Intel and AMD systems was also dependent on your GPU of choice. 

The pairing with a high-end GPU (e.g. the GTX 1080's, 1080 Ti's and 2080 Ti's nowadays) will only amplify the difference. That is totally understandable, you can only review a CPU by actually looking at the highest tier of performance, but with more commodity systems (with a GTX 1060 level of performance) will show less of a difference.

Maybe they meant top tier performance in that way?
Would still be an exaggeration.. They offer good value for the money and move the industry forward with more cores though!

 

 

[spartaman64]

yes because gaming is the only thing people use cpus for right?

[JoeCoke]

1 hour ago, BlueGoliath said:

Why punish me by locking the thread instead of removing their troll comments and giving them a warning?  You still have people attempting to derail it by suggesting there is something wrong with my setup

I know it's tempting to call people who disagree with you or question you trolls, but no one here is "trolling". It's hard to derail a thread when it was never on the rails. The whole point of this thread is to accuse LTT of flip flopping, while simultaneously making claims such as a 1800x can't achieve a stable 60 in GTAVO. You say us bringing up an issue with your computer is derailing, when you are the one who brought your system into the conversation. And then when your own examples (using your system) are brought into question, you beg a mod publicly to remove those criticisms. I'm done posting here now, this thread is a joke.

[Zando_]

1 minute ago, 2Buck said:

while simultaneously making claims such as a 1800x can't achieve a stable 60 in GTAVO. 

Lmao, pretty sure people have pushed a decent fps in GTAV with the same dual Xeon setup first gen Mac Pros use (Core 2 era quad core Xeons for 8c/8t total). 

[BlueGoliath]

7 hours ago, Sauron said:

If you're going to complain about it dipping to 58 once every 5 hours you don't have your priorities in order. Oh, and you backed off from what was a straight up lie

 

yes, the 1800x can ABSOLUTELY achieve stable 60 in gta v.

 

 

The benchmark isn't representative of actual in-game performance. How many times has this been said now? Do you even play GTA 5 on a first generation Ryzen CPU?

 

 

Quote

 

Yes, it was. It just wasn't the performance king.

 

It isn't even close in most games. Linus even said in his second generation review that Intel was better. Since when is a 50 FPS gap  insignificant? 

 

 

Quote

Because your using "your setup" as be all end all to your argument to the 1800X, when it's extremely easy to disprove.

 

Yeah, OK. It isn't like AIDA has an 1800x with 80ns of memory latency as part of its default benchmark listings.

 

 

 

Something tells me noone here has ever bought or used a first generation Ryzen CPU...

 

Quote

 

yes because gaming is the only thing people use cpus for right?

 

 

Ah, you mean the "creator" buzzword that tech outlets like LTT helped bring into popularity by putting a crapload emphasis on things that don't reallymatter to make Ryzen look better and motherboard manufacturers are now attempting to capitalize on by releasing "creator" motherboards.

 

There is no denying it's a market and part of LTT's audience but uh... it isn't exactly a huge market.

 

 

[Drak3]

15 minutes ago, BlueGoliath said:

The benchmark isn't representative of actual in-game performance. How many times has this been said now?

Not enough to make it true.

 

16 minutes ago, BlueGoliath said:

It isn't even close in most games.

Yes, it was. And is.

 

16 minutes ago, BlueGoliath said:

It isn't like AIDA has an 1800x with 80ns of memory latency as part of its default benchmark listings.

 

Except that 80ns isn't making your bad claim good, ergo it's irrelevant.

 

Also, I like how you stopped actually quoting me to avoid my responces.

[Sauron]

6 minutes ago, BlueGoliath said:

The benchmark isn't representative of actual in-game performance. How many times has this been said now? Do you even play GTA 5 on a first generation Ryzen CPU?

I prefer empirical testing over anecdotal evidence, thanks.

6 minutes ago, BlueGoliath said:

Yeah, OK. It isn't like AIDA has an 1800x with 80ns of memory latency as part of its default benchmark listings.

Irrelevant to the conversation.

7 minutes ago, BlueGoliath said:

Something tells me noone here has ever bought or used a first generation Ryzen CPU...

Have you bought and used the 7700k you keep praising? No? Then I guess nothing you say holds any water, after all by your own logic if you haven't tested it yourself you just can't know.

9 minutes ago, BlueGoliath said:

Ah, you mean the "creator" buzzword that tech outlets like LTT helped bring into popularity by putting a crapload emphasis on things that don't reallymatter to make Ryzen look better and motherboard manufacturers are now attempting to capitalize on by releasing "creator" motherboards.

Nah, they mean a very real and widespread use case for high end CPUs. Computers aren't just toys, people work on them. I could mention a boatload of workloads where having more cores is a huge advantage, but you'd probably just dismiss them because you personally don't need them. Not to mention the price, heat output and power consumption advantage AMD holds across the board.

12 minutes ago, BlueGoliath said:

There is no denying it's a market and part of LTT's audience but uh... it isn't exactly a huge market.

Huh, is that something you can back up with a statistic or anything other than your gut feeling? Do you think the entirety of the media production industry is a small niche? Do you think the ads you see every day in the thousands just spontaneously make themselves? What about movies, tv series, cartoons, youtube videos? The games you think are such a huge part of the marketshare that everything else must come second?

[BlueGoliath]

7 hours ago, 2Buck said:

I know it's tempting to call people who disagree with you or question you trolls, but no one here is "trolling". It's hard to derail a thread when it was never on the rails. The whole point of this thread is to accuse LTT of flip flopping, while simultaneously making claims such as a 1800x can't achieve a stable 60 in GTAVO. You say us bringing up an issue with your computer is derailing, when you are the one who brought your system into the conversation. And then when your own examples (using your system) are brought into question, you beg a mod publicly to remove those criticisms. I'm done posting here now, this thread is a joke.

 

The first few comments where nothing but mockery and derailing as if none of it mattered. What exactly was that kit-kat image?

 

The thread was perfectly on rails and GTA 5 was only brought up later as a counter argument that there was no "update" to *anything* that massively improved Ryzen performance and made it as good or near as good as a 7700k. There was no proof of such claim and he left since he couldn't back up his claim. 

 

(Side note, there was some thread scheduler changes in Windows 10 that was supposed to boost performance but I ain't seeing it... and i'm in the Insider Preview).

 

This thread has been nothing but derailing and trolling since it was created with few exceptions like @Minibois and one or two others who respectful disagreed and I appreciate it. Community standards went out the window practically with the first reply.

 

The amount of people here who think they know what they are talking about despite not owning or using a Ryzen 1800x(or second generation, for that matter) is pretty insane. Even more so is the amount of people here who think benchmarks are always indicative of in-game performance, especially in an online setting. I ain't claiming I know everything but given that Cinebench reports a 3933 score(well above the stock 1700x default listing) and 78ns in AIDA compared to the default listing of 80ns at 2933Mhz.... There is nothing wrong with my system. It's all within margin of error or can be explained by GPU differences(1080 vs 1080 TI).

 

edit: added the score. Actually better than 3922 after a rerun...

[BlueGoliath]

7 hours ago, Drak3 said:

 

 

 

Those two statements don't conflict with each other. A 5 FPS gain at 60 fps isn't the same as 5 FPS at 300 FPS in terms of required GPU performance to achieve it. The amount of work per frame decreases as FPS increases assuming hardware can keep up...

[Drak3]

Just now, BlueGoliath said:

Those two statements don't conflict with each other.

Given that that was your responce to the 1800X handily doing DOUBLE what you claimed it couldn't,, I'd say that yes, they do by your logic.

[BlueGoliath]

7 hours ago, Drak3 said:

Given that that was your responce to the 1800X handily doing DOUBLE what you claimed it couldn't,, I'd say that yes, they do by your logic.

 

Where is it doing 120FPS? I only see about 100FPS and barely 60 FPS in the 99th percentile...

[Drak3]

8 minutes ago, BlueGoliath said:

 

Where is it doing 120FPS? I only see about 100FPS and barely 60 FPS in the 99th percentile...

My link. Averaging 140FPS in GTA V with a 1080TI. Low point was 120.

 

Which anyone with a Zen 1 hexacore+ should expect.

[Guest]

13 hours ago, BlueGoliath said:

 

Sure, not everyone plays CS:GO but the fact that the 1800x can't even achieve stable 60FPS in GTA 5 online is kind of a big deal. In some games it doesn't matter that much while in others it does.

 

Still, the irony of an AMD CPU being outperformed by a older Intel one is ironic.

 

 

Why would I be an Intel fanboy if I had an AMD CPU? Hello?

My 1800X can achieve 60fps in GTAV don't know what you're on about. Try adjusting your settings. 

[BlueGoliath]

7 hours ago, Drak3 said:

My link. Averaging 140FPS in GTA V with a 1080TI. Low point was 120.

 

If the previous chart showed the 1800x w/ GTX 1080 could barely keep above 60 FPS how the actual hell can anyone get 120? That benchmark was probably someone doing extreme memory OCing and disabled SMT. I had gotten similar insane performance while doing just that before...

 

This random YouTube video even shows dips into the 70s.

 

edit: here is another one, but this time in online. Dips into the 50s.

 

edit 2: here is the userbenchmark which represents my hardware. No 120 FPS here.

 

Quote

Which anyone with a Zen 1 hexacore+ should expect.

 

Ironic.

 

I have to ask, do you even own a Ryzen CPU? Because you seem to just be pulling from benchmarks from the internet and have no actual personal experience.

[Drak3]

49 minutes ago, BlueGoliath said:

This random YouTube video even shows dips into the 70s

The 7700K can have dips into the 20's. But that's only a sliver of time that users don't notice.

 

And dipping to 70 is still disproving your claim.

50 minutes ago, BlueGoliath said:

do you even own a Ryzen CPU?

Irrelevant. Empirical evidense has proved you wrong multiple times at this point.

 

You dismissing a benchmark becauae it's not representative doesn't work either, because the GTAV benchmark exists for the sole reason to see if a system can run GTAV and at what settings. What the benchmark does is pretty damn close to the game does. Margin of error differences.