Ryzen 3rd Gen vs Intel 9th Gen

[Varus]

Switched the PSU. @GoldenLag The GPU is going to be used for 1080p gaming and rendering so would like to keep it under $300. I was thinking something like a 1660 ti but was also looking at the rtx 2060 super but at $400 I am not sure if it's worth it.

[Fasauceome]

Just now, The Torrent said:

Why is it wrong? Effective speed maybe wrong but benchmarks are still benchmarks.

"Benchmarks" are not all the same. If you took the time to watch the video I shared, you'd learn that the userbenchmark data is not collected in a controlled environment, contaminating every single statistic taken. Their data is wholly unreliable.

[The Torrent]

1 minute ago, GoldenLag said:

The ex920 is NVMe. I woudlnt bither overpaying for a more expencive, while also worse drive.....

Then why reffer to userbenchmark? And not like reviewers?

I never read any reviews reviews are long and it’s a busy life benchmarks are quick and to the point. Reviews depend on too much benchmarks are loads of examples put together and no possible way of bias.

 

any cpu company can go buy a review I’d easily be bought.

[5x5]

Just now, The Torrent said:

No.

 

the effective speed is drivel but no one is telling u to live by that.

 

no insult. I simply said your pissing me off and you can’t read as it’s the third time I’m making the above point and your still not acknowledging it.

 

Telling me that I am illiterate is an insult. And no, userbenchmarks other fiwlds are also senseless drivel as GN explained thoroughly. It's bad faith even telling som one to look at it, let alone basing a purchase on it. Please provide reputable datasets from trusted sources or don't give bad advice and waste other people's money.

[The Torrent]

Just now, fasauceome said:

"Benchmarks" are not all the same. If you took the time to watch the video I shared, you'd learn that the userbenchmark data is not collected in a controlled environment, contaminating every single statistic taken. Their data is wholly unreliable.

But there’s loads of benchmarks. It’s not 1 or two there’s many rigs tested I’ve even done 3 of mine.

[Fasauceome]

Just now, The Torrent said:

But there’s loads of benchmarks. It’s not 1 or two there’s many rigs tested I’ve even done 3 of mine.

That's my point, too many test rigs with uncontrolled setups. The data is all over the place, from any average Joe, making it unreliable.

[5x5]

Just now, The Torrent said:

But there’s loads of benchmarks. It’s not 1 or two there’s many rigs tested I’ve even done 3 of mine.

It's not a controlled environment therefore it has no scientific or logistical vsluem it. Is. Drivel.

[The Torrent]

Just now, fasauceome said:

That's my point, too many test rigs with uncontrolled setups. The data is all over the place, from any average Joe, making it unreliable.

Find the median.

 

thats how irl works, that’s how they do real irl survey’s. Range don’t matter.

[GoldenLag]

2 minutes ago, The Torrent said:

any cpu company can go buy a review I’d easily be bought.

Sure they can.

 

Can they buy a review from someone reputable who have their entire living based ln the fact they are reputable? Nope. Hardware unboxed and GN and LTT are a thing. You cant buy them, you simply cant. 

[5x5]

Just now, The Torrent said:

Find the median.

 

thats how irl works, that’s how they do real irl survey’s. Range don’t matter.

It absolutely does. This is not gender studies. This is cold hard data. Manipulating it is basically lying to yourself and others.

[Fasauceome]

Just now, The Torrent said:

Find the median.

 

thats how irl works, that’s how they do real irl survey’s. Range don’t matter.

this isn't a survey, it's a technological examination of quantifiable performance and data, completely different.

 

3 minutes ago, The Torrent said:

I never read any reviews reviews are long and it’s a busy life benchmarks are quick and to the point. Reviews depend on too much benchmarks are loads of examples put together and no possible way of bias.

The reason reviews are more informative than random benchmarks is because they explain their testing methodology to be understood and extrapolated on. They also include benchmarks within themselves, so your reasoning is faulty.

[airdeano]

look, jamming up someones thread with this credulous back-n-forth with no real outcome is non-productive. please be respectful of the community's needs in support and suggestions. if you cannot abide, then do not lend that sage advice.

<-UNLOCKED->

[hello_there_123]

46 minutes ago, fasauceome said:

this is pretty crap. Aim for the Corsair CX550 or 550M instead, slightly higher price tag but it goes a long way.

https://pcpartpicker.com/product/FdRFf7/corsair-cx-2017-550w-80-bronze-certified-atx-power-supply-cp-9020121-na

The CX450 and 450M are similar price for better quality but only have one 8 pin PCIe connector.

Confuse as to why one should get cx550 , when formula gold 550 , much higher quality is on Newegg , for just $5 more...

 

46 minutes ago, GoldenLag said:

Swao for a ex920 1TB. Its faster and overall better drive. 

When sabrent rocket , is $8 cheaper??

[Fasauceome]

4 minutes ago, hello_there_123 said:

Confuse as to why one should get cx550 , when formula gold 550 , much higher quality is on Newegg , for just $5 more...

oh crap I forgot the price is holding so low these days

[GamerDude]

2 hours ago, The Torrent said:

Userbenchmark i9 9900k vs 3900x

FFS, NEVER EVAH use Userbenchmark, it's FUBAR'ed beyond belief. As for which CPU, a 3700X/3600X would be good enough to last for a good long while, while not being quite up to Intel's equivalent (price-wise), they're no slouch when it comes to gaming....sure, there's a gap in framerate when it comes to gaming, but when you go up in resolution, that gap converge, or at the very least, that gap narrows a fair bit. For productivity where hyper-threading (or SMT as AMD calls it) is a great benefit, the Ryzen 3xxx series shines.

[bignaz]

I have no problem with my 3900x and gaming. 

 

Note cpuz on my system is having issues reporting the right voltage at times. That's 1.38v  

 

 

But seriously don't expect this.  And even if your cpu can pull it what by the looks of it less then 1% can even hit 4.6 all core.  You still have to cool it and it makes a ton of heat. 

 

[paulmohr]

Userbenchmark isn't even a benchmark suite. It is of no valid use in comparing hardware. You might as well go by the user reviews on amazon, actually they might be more accurate.

 

Quoting that site is like saying "My brother in law is a mechanic and his six year old son told me ford is better than chevy, so I am going to buy a ford.".

 

Why even comment on something you didn't even bother to read? It took me all of about a minute to read that post. You must have a seizure when you read one of my posts lol.

 

As for the original post, I think most of the others have pointed you in a good direction. Some of those reviews have actually tested those cpus on those programs. Look at the data and decide if the performance gain or loss is worth the price, or going with a brand you like better. If someone wants to spend a little more on a brand they like or trust more for whatever reason I don't knock them for it. Or take a slight performance hit for less money, either way. Honestly if Apple released the most awesome fastest computer in the world I still wouldn't buy it. I just hate Apple that much lol.

[GoldenLag]

8 hours ago, hello_there_123 said:

When sabrent rocket , is $8 cheaper??

Iirc the sabrent rocket is cacheless.

 

So i dont see why you would get the sabrent rocket. 

[GoldenLag]

7 hours ago, bignaz said:

But seriously don't expect this.  And even if your cpu can pull it what by the looks of it less then 1% can even hit 4.6 all core

The real question is if you are actually getting the performance of 4,6 ghz. Zen has allways been picky about this. Even if you run seemingly 4,4 ghz on a 2700x, it would be worse than a stock 2700 in terms of performance.

[Stormseeker9]

10 hours ago, The Torrent said:

I never read any reviews reviews are long and it’s a busy life benchmarks are quick and to the point.

There’s your problem 

[The Torrent]

12 minutes ago, Stormseeker9 said:

There’s your problem 

My problem that benchmarks are quick and to the point?

 

how is thst a problem?

 

 

i thought this discussion was closed.

[bignaz]

2 hours ago, GoldenLag said:

The real question is if you are actually getting the performance of 4,6 ghz. Zen has allways been picky about this. Even if you run seemingly 4,4 ghz on a 2700x, it would be worse than a stock 2700 in terms of performance.

Yeah. At 4.7+ it wrecks my i9.  But as of right now there's only maybe 2 that can pull this all core and maybe a few more that can do it on limited cores.  And it's not easy getting here even if your cpu can do it.

[hello_there_123]

5 hours ago, GoldenLag said:

Iirc the sabrent rocket is cacheless.

 

So i dont see why you would get the sabrent rocket. 

https://www.tomshardware.com/reviews/corsair-force-mp510-ssd,5848.html

 

Whoever told you , that the sabrent rocket is cacheless clearly has no idea what they,re talking about...

[GoldenLag]

34 minutes ago, hello_there_123 said:

https://www.tomshardware.com/reviews/corsair-force-mp510-ssd,5848.html

 

Whoever told you , that the sabrent rocket is cacheless clearly has no idea what they,re talking about...

It has a cache, but a much smaller one.

 

You can also save some money by going with the Crucial P1 and 660p from intel. 

[hello_there_123]

10 minutes ago, GoldenLag said:

It has a cache, but a much smaller one.

That wasn,t the point of matter , you said that it didnt have a cache at all which was clearly false. 

And i,ll just leave this here... https://old.reddit.com/r/buildapcsales/comments/bg7ttw/ssd_hp_ex920_m2_1tb_13591_159892398_w_save15_at/elj98ql/