[Hardware Unboxed] R7 1700 @4Ghz + 3200Mhz RAm = ~7820X @4.5Ghz

[Actaeon]

Yep, TechSpot again. His results with SKL-X is consistently lower than every other review. Nothing to see here, move along.

 

[moonbogg]

I'm glad he used the R7 1700 for this test instead of any other Ryzen chip. The 1700 is a total menace. I see people mention that Intel chips are still superior, but when the performance is this close and the price is so different, it truly makes the Intel offering completely DOA. This is AMD's time to really shine since Intel pulled a Bulldozer with Skylake-X by ramping power and temps through the roof while being clever enough to actually go backwards in performance. That takes effort.
AMD wins. They just plain win. Its that simple.

 

[IndyColtsFan]

And this is precisely why my main build comes down to Threadripper or Coffee Lake.

 

[TahoeDust]

I see people mention that Intel chips are still superior, but when the performance is this close and the price is so different, it truly makes the Intel offering completely DOA.

My 7820x and MOBO was $500 more than the 1700 and mobo I looked at. The CPU performance is ~20% from the tests I have seen and done. $500 for 20%... The value of that 20% is going to really be budget dependent.

 

[IndyColtsFan]

My 7820x and MOBO was $500 more than the 1700 and mobo I looked at. The CPU performance is ~20% from the tests I have seen and done. $500 for 20%... The value of that 20% is going to really be budget dependent.

I thought it was closer to 10-20% difference on average. Nevertheless, $500 is a big chunk of change that could be used on other components - better/large drives, better graphic cards, faster RAM, etc. It's why I'm strongly considering a 1950X coupled with the fastest graphic card I can find - sure, many Intel chips would beat it at gaming but am I going to care if I get 150 fps instead of 170? Probably not, and that tradeoff buys me more cores for the money.

 

[lobz]

The 1900X has fundamentally the same problem as Kaby Lake-X - that is to say, if you're already buying in to a high-end platform like X399 or X299, your CPU budget is probably higher. And if your CPU budget isn't higher, you should probably save money by going for the mainstream motherboard and then have more money to spend on things like a better GPU or more RAM or SSD capacity (depending on your intended use).

wrong. KL-X offers practically nothig over KL-S, just 100 MHz higher base clock and a higher TDP. The 8 core TR on the other hand, well..... offers 200MHz more base clock, double the mem channels, more than double the usable PCIe lanes etc... wrong comparison in my eyes

 

[lobz]

gcc segmentation faults on Ryzen / Linux7 https://community.amd.com/thread/215773?start=555&tstart=0
New Ryzen FMA bug https://community.amd.com/thread/218643

That is on top of what AMD already fixed on Ryzen (FMA3 bug, VME bug, multiple AGESA releases until memory finally works at high speed).

many issues = beta product released too early

If AMD fully fixes their product I may consider their CPUs.

this is one of the most ignorant posts I've read in a while

 

[TahoeDust]

I thought it was closer to 10-20% difference on average. Nevertheless, $500 is a big chunk of change that could be used on other components - better/large drives, better graphic cards, faster RAM, etc. It's why I'm strongly considering a 1950X coupled with the fastest graphic card I can find - sure, many Intel chips would beat it at gaming but am I going to care if I get 150 fps instead of 170? Probably not, and that tradeoff buys me more cores for the money.

What if you didn't need to save that $500 on the CPU/MOBO in order to buy whatever drives, graphics cards, and RAM you wanted or needed? Some enthusiasts prefer performance over value. In a high end $3500 build, saving $500 only lowers the price by 16%....the price difference is now pretty damn close to the performance difference.

 

[lobz]

My 7820x and MOBO was $500 more than the 1700 and mobo I looked at. The CPU performance is ~20% from the tests I have seen and done. $500 for 20%... The value of that 20% is going to really be budget dependent.

That's actually the difference between a mid range and a high-end VGA plus better RAM. It counts a lot, depending on your use case.

 

[lobz]

AMD's product tiers are very well positioned and really show Intel's faults there imho.

This. Zen execution is almost perfect from AMD's side, they should just have work much better together with the mobo manufacturers.

 

[lobz]

What if you didn't need to save that $500 on the CPU/MOBO to buy whatever drives, graphics cards, and RAM you wanted or needed?

then you probably don't compare these 2 specific CPUs at all.

edit: I've just read your later posts, and while I find your case a bit unconventional, you're right of course

 

[Malogeek]

then you probably don't compare these 2 specific CPUs at all.

Well you should still do your homework and figure out which one is best for your specific use cases.

 

[TahoeDust]

then you probably don't compare these 2 specific CPUs at all.

edit: I've just read your later posts, and while I find your case a bit unconventional, you're right of course

Sorry. I edited the post that you quoted to clarify. It now reads as follows...

What if you didn't need to save that $500 on the CPU/MOBO in order to buy whatever drives, graphics cards, and RAM you wanted or needed? Some enthusiasts prefer performance over value. In a high end $3500 build, saving $500 only lowers the price by 16%....the price difference is now pretty damn close to the performance difference.

 

[IEC]

What if you didn't need to save that $500 on the CPU/MOBO in order to buy whatever drives, graphics cards, and RAM you wanted or needed? Some enthusiasts prefer performance over value. In a high end $3500 build, saving $500 only lowers the price by 16%....the price difference is now pretty damn close to the performance difference.

In my case you spring for a Threadripper 1950X, X399 board, and tons of RAM. Because it's only $400 more than a i7-7820X but has DOUBLE the cores.

 

[IndyColtsFan]

What if you didn't need to save that $500 on the CPU/MOBO in order to buy whatever drives, graphics cards, and RAM you wanted or needed? Some enthusiasts prefer performance over value. In a high end $3500 build, saving $500 only lowers the price by 16%....the price difference is now pretty damn close to the performance difference.

If you asked me 10-15 years ago, I'd probably spend the $500. If you asked me now, I'd save the $500 and apply it to my other hobbies.

$500 isn't much money to me and if I really wanted something and the cost was $500, I'd buy it. However, as I've gotten older, I have started investing in other interests and more and more consider bang for the buck. The 7820x is a really tough proposition IMO given the total platform cost and where it sits in relation to the 1920x.

 

[IndyColtsFan]

In my case you spring for a Threadripper 1950X, X399 board, and tons of RAM. Because it's only $400 more than a i7-7820X but has DOUBLE the cores.

And the 1920X has 50% more cores for only $200 more than the 7820x.

 

[TahoeDust]

In my case you spring for a Threadripper 1950X, X399 board, and tons of RAM. Because it's only $400 more than a i7-7820X but has DOUBLE the cores.

Definitely not a bad option. Honestly, I still game on my machine a lot, and the anticipated low clock speeds and single core performance kept me from holding out and waiting for TR to become an option. When the head to head gaming comparisons come out we'll see.

 

[IEC]

Definitely not a bad option. Honestly, I still game on my machine a lot, and the anticipated low clock speeds and single core performance kept me from holding out and waiting for TR to become an option. When the head to head gaming comparisons come out we'll see.

Gaming is a minor usage for me, and at 4K I doubt I would be able to tell a difference between a TR 1950X and *any* Intel processor in gaming.

 

[krumme]

What if you didn't need to save that $500 on the CPU/MOBO in order to buy whatever drives, graphics cards, and RAM you wanted or needed? Some enthusiasts prefer performance over value. In a high end $3500 build, saving $500 only lowers the price by 16%....the price difference is now pretty damn close to the performance difference.

Why not get a 1950 then?
It creams a 7820x. As cl 6c is the gamers 144Hz cpu anyway.

The 7820x plus mb is more than twice the cost of 1700 plus 350 or cheap 370. ~120%. Not remotely same segment imo.

I would say 1950 and 7820x caters to same power users even if the 1950 total cost is like 40% more.

The comparison should be 1950 vs 7820x and 7900x.

Everyone and his brother gets 1950 now or 6c cl or they go 1600(x)/1700.

To be frank after tr i only think 8700k and low end pentium is where Intel is best.

Intel better get in motion because 7nm high perf ibm process is going to make a wreck on the intel line with zen2. It will make the soi athlon times come live again. How ironic.

Its incredible this dog slow lpp 14nm can even make such an impact on desktop space. I never ever expected that even remotely.

 

[Anarchist Mae]

Yep, TechSpot again. His results with SKL-X is consistently lower than every other review. Nothing to see here, move along.

Compared to which reviews?

 

[Blockheadfan]

Yep, TechSpot again. His results with SKL-X is consistently lower than every other review. Nothing to see here, move along.

Care to show any tests run in an identical way that contradict his results? The only nothing I've seen here is data to back up what you've claimed.

 

[Crumpet]

Compared to which reviews?

That would be a WCCFtech test where SK-X has a 1000Mhz RAM advantage, a Eurogamer test where there may be a $600 Mhz advantage, but their testing suite is incomplete or flawed and doesn't mention the Ryzen setup, and the last was from Hardware.info where they ran all systems at 2666, which whilst I understand the logic behind doing so, we were discussing Ryzen + fast RAM gaming performance, and the known FACT that at such a low RAM speed Intel was going to walk all over Ryzen.

So, in my opinion, it means compared to 3 invalid reviews.

 

[coercitiv]

So, in my opinion, it means compared to 3 invalid reviews.

No, it doesn't mean the other reviews are invalid, it means we cannot compare review outcomes based on apples and oranges.

A review that pairs 2666 RAM with Zen is just as useful to consumers planning to buy 2666 RAM as the review using 3200 C14 RAM is for those looking to invest in high performance RAM.

Both Digital Foundry and Hardware Unboxed are excellent reviewers in my opinion, as long as people read their reviews and assess based on context rather than trying to pick and choose results to fit the all mighty agenda.

 

[krumme]

No, it doesn't mean the other reviews are invalid, it means we cannot compare review outcomes based on apples and oranges.

A review that pairs 2666 RAM with Zen is just as useful to consumers planning to buy 2666 RAM as the review using 3200 C14 RAM is for those looking to invest in high performance RAM.

Both Digital Foundry and Hardware Unboxed are excellent reviewers in my opinion, as long as people read their reviews and assess based on context rather than trying to pick and choose results to fit the all mighty agenda.

I tend to disagree.
If you benchmark with games in focus not using 3200 for zen is flawed. At the same time it needs to be added to the cost as well as mb cost differences.

 

[Crumpet]

No, it doesn't mean the other reviews are invalid, it means we cannot compare review outcomes based on apples and oranges.

A review that pairs 2666 RAM with Zen is just as useful to consumers planning to buy 2666 RAM as the review using 3200 C14 RAM is for those looking to invest in high performance RAM.

Both Digital Foundry and Hardware Unboxed are excellent reviewers in my opinion, as long as people read their reviews and assess based on context rather than trying to pick and choose results to fit the all mighty agenda.

My problem is when a review pairs Ryzen with 2666Mhz RAM to compare it against a competitor running 3600Mhz RAM, and then pointing out the poorer gaming performance.