Intel bug incoming? Meltdown and Spectre exploits

gavinh87 · 10 Jan 2018 at 17:11

humbug said:
If you're that bothered you would upgrade to a Ryzen 2600 and still save money vs getting the 8700K, that's how much more expensive the 8700K is and its about to because a bit null with the Ryzen 2 about to launch.

As for your slides anyone can cherry pick one slide and try to make out its because a lack of overall performance, look how the low end Ryzen 1500X 4 core trashes the top on the range Intel 6 core in this game..

We've already been though this picture.
Are you saying that civ6 uses some secret sauce as that is the only single game where the r5 can out do an i7. Or maybe, just maybe the reviewers are testing it wrong. Just as Steve from gamers nexus said.
Anyhow, buying an 1600 and then a 2600 isn't cheaper and you'll still be stuck with games that just don't like AMD.
You're all aboard this ryzen refresh train, are you going to put your hand in your pocket this time round?

CAT-THE-FIFTH · 10 Jan 2018 at 17:12

pete910 said:
Good to know. I may just give PS2 another bash If I can be arsed to boot win, Hate to think the hours I might have in it. I jumped on that on release which seems like an age ago now!

Its still a fun game and very few online games have that sense of scale for me,plus the Free To Play model they have is fair IMHO OFC.

humbug · 10 Jan 2018 at 17:17

gavinh87 said:
We've already been though this picture.
Are you saying that civ6 uses some secret sauce as that is the only single game where the r5 can out do an i7. Or maybe, just maybe the reviewers are testing it wrong. Just as Steve from gamers nexus said.
Anyhow, buying an 1600 and then a 2600 isn't cheaper and you'll still be stuck with games that just don't like AMD.
You're all aboard this ryzen refresh train, are you going to put your hand in your pocket this time round?

civ6 uses some secret sauce

just maybe the reviewers are testing it wrong

lol..... yeah, right. now your defensiveness is getting ridiculous

Its deliberately designed to push the CPU to its limits, to pushed AI Math, Physics and Ray Tracing as hard and as fast as the CPU is capable off, that's the "secrete source".

CAT-THE-FIFTH · 10 Jan 2018 at 17:20

OTH,I suppose at least was good for Intel to pull the Coffee Lake launch forward - imagine if they had launched it now into this *********!

humbug · 10 Jan 2018 at 17:22

CAT-THE-FIFTH said:
OTH,I suppose at least was good for Intel to pull the Coffee Lake launch forward - imagine if they had launched it now into this *********!

Certainly review benchmarks had they been done now would look quite different.

gavinh87 · 10 Jan 2018 at 17:27

humbug said:
lol..... yeah, right. now your defensiveness is getting ridiculous

Its deliberately designed to push the CPU to its limits, to pushed AI Math, Physics and Ray Tracing as hard and as fast as the CPU is capable off, that's the "secrete source".

Gamers nexus explained the reason for this result. Believe it or ignore it. It is one single game out of many where AMD cannot keep up. Those are facts. Those people on ryzen now will be upgrading sooner than those on a i7 coffelake will be. It's false economy.
It will be interesting to see how bad ryzen will bottleneck the next gen GPUs.

pete910 · 10 Jan 2018 at 17:29

CAT-THE-FIFTH said:
Its still a fun game and very few online games have that sense of scale for me,plus the Free To Play model they have is fair IMHO OFC.

Off topic I know but remembered my acc login for it and thought I'd look !

Code:

Total Time Played 21d 11h

humbug · 10 Jan 2018 at 17:32

gavinh87 said:
Gamers nexus explained the reason for this result. Believe it or ignore it. It is one single game out of many where AMD cannot keep up. Those are facts. Those people on ryzen now will be upgrading sooner than those on a i7 coffelake will be. It's false economy.
It will be interesting to see how bad ryzen will bottleneck the next gen GPUs.

Gamers Nexus DO NOT disagree with reviewers results, they did not say reviewers are testing it wrong, you're straw clutching.

It will be interesting to see how bad ryzen will bottleneck the next gen GPUs.

You need to take a break with this furious hate, you're starting to look like you're boiling your own head over hardware vendors.

gavinh87 · 10 Jan 2018 at 17:36

humbug said:
I know it, i have seen it, Gamers Nexus DO NOT disagree with reviewers results, they did not say reviewers are testing it wrong, you're straw clutching.

You need to take a break with this furious hate, you're starting to look like you're boiling you're own head over hardware vendors.

Take a look again, he actually explains the fps is high because it's basically idle.

I'm not hating anything. What I am saying ryzen is not future proof. It's showing it's limits with gpus already. And upgrading later on is false economy.
I'm out of this thread before you report me again.

CAT-THE-FIFTH · 10 Jan 2018 at 17:38

pete910 said:
Off topic I know but remembered my acc login for it and thought I'd look !

Code:

Total Time Played 21d 11h

I think I might have a few hundred hours in the game too!!

humbug · 10 Jan 2018 at 17:42

gavinh87 said:
Take a look again, he actually explains the fps is high because it's basically idle.

I'm not hating anything. What I am saying ryzen is not future proof. It's showing it's limits with gpus already. And upgrading later on is false economy.
I'm out of this thread before you report me again.

No, You don't get higher Frame Rates with the same GPU if the CPU is weaker, if its idling the performance is the same across the board provided the GPU is the same performance.

Effectively what you are saying is the CPU performance levels are inverted, the higher the FPS are 'on the same GPU' the slower the CPU's. its a completely bonkers argument.
You only get that level of idiocy in arguments if one is straw clutching to the extreme.

DragonQ · 10 Jan 2018 at 17:43

humbug said:
If you're that bothered you would upgrade to a Ryzen 2600 and still save money vs getting the 8700K, that's how much more expensive the 8700K is and its about to because a bit null with the Ryzen 2 about to launch.

As for your slides anyone can cherry pick one slide and try to make out its because a lack of overall performance, look how the low end Ryzen 1500X 4 core trashes the top on the range Intel 6 core in this game..

We don't know anything concrete about Ryzen+'s performance yet.

humbug · 10 Jan 2018 at 17:51

DragonQ said:
We don't know anything concrete about Ryzen+'s performance yet.

We do know Ryzen + is built on Samsung's 12nm, which has "15% higher circuits density" and "more than a 10% performance improvement"

So we can say at least a 10% uplift in performance just from moving from the current node to that one, that's without any architectural tweaks, just a straight die shrink, the + 15% density can go to adding to the architecture or saving money.

gavinh87 · 10 Jan 2018 at 17:59

humbug said:
No, You don't get higher Frame Rates with the same GPU if the CPU is weaker, if its idling the performance is the same across the board provided the GPU is the same performance.

Effectively what you are saying is the CPU performance levels are inverted, the higher the FPS are 'on the same GPU' the slower the CPU's. its a completely bonkers argument.
You only get that level of idiocy in arguments if one is straw clutching to the extreme.

Civilization VI is a frequency-intensive benchmark, and has proven that turn time is less dependent on cores. Note also that the Civ AI benchmark should not be used to test FPS, as worse CPUs will score higher framerates as a result of spending more time on static screens.

Somewhat unsurprisingly, the i3-8350K performs about where the R5 1600X (1600 non-X on Amazon here) at 4.1GHz performs, showing the frequency importance for this title. The 8350K completes each turn in 18 seconds, totaling 90 seconds for all 5 turns. The R5 1600X finishes in 17.9 seconds at 4.1GHz, or 19.2 seconds stock. Intel’s i5-8400 completes the turns in roughly 17.5 seconds, depending on memory frequency.

Overclocking the 8350K to 4.8GHz reduces turn time requirements by 9.6%, tying the CPU with the stock 8700K.

This is how civ6 should be benchmarked. It falls right in line with how we've come to see ryzen perform against intel in gaming.
As for ryzen refresh that 10% came from a comparison against a 16nm node, not the current 14nm.

DragonQ · 10 Jan 2018 at 18:00

humbug said:
We do know Ryzen + is built on Samsung's 12nm, which has "15% higher circuits density" and "more than a 10% performance improvement"

So we can say at least a 10% uplift in performance just from moving from the current node to that one, that's without any architectural tweaks, just a straight die shrink, the + 15% density can go to adding to the architecture or saving money.

Yes that's what the GF slides say (along with claiming the node is built for 5 GHz parts compared to ~3 GHz parts) but Intel also claimed a 15% performance improvement from Skylake to Kaby Lake in their promotional slides, so I think we should be sceptical until we see actual numbers from AMD. I am hoping for a 10+% uplift in clock speeds but we'll see.

humbug · 10 Jan 2018 at 18:09

gavinh87 said:
This is how civ6 should be benchmarked. It falls right in line with how we've come to see ryzen perform against intel in gaming.
As for ryzen refresh that 10% came from a comparison against a 16nm node, not the current 14nm.

Turn time depends on a pot pourri of different things, every reviewer likes to say they do it best but the fact is turn time is just one of many ways Civ6 uses hardware, the problem with measuring turn times every turn is different to the last, its like going on to a full 64 player BF4 server and claiming each run is the same, that's not possible, you can get them close if you're very very careful but you can never account for the other 64 variables.

pete910 · 10 Jan 2018 at 18:11

Stay mostly on topic guys please!

gavinh87 · 10 Jan 2018 at 18:13

humbug said:
Turn time depends on a pot pourri of different things, every reviewer like to say they do it bets but the fact is turn time is just one of many ways Civ6 uses hardware, the problem with measuring turn times every turn is different to the last, its like going on to a full 64 player BF4 server and claiming each run is the same, that's not possible, you can get them close if you're very very carful but you can never account for the other 64 variables.

Its a canned benchmark.....

humbug · 10 Jan 2018 at 18:16

Not that bit it isn't ^^^^

DragonQ said:
Yes that's what the GF slides say (along with claiming the node is built for 5 GHz parts compared to ~3 GHz parts) but Intel also claimed a 15% performance improvement from Skylake to Kaby Lake in their promotional slides, so I think we should be sceptical until we see actual numbers from AMD. I am hoping for a 10+% uplift in clock speeds but we'll see.

The actual performance of Ryzen 1 ended up significantly higher than AMD and Glofo claimed.

Of course they could be lying this time round but Ryzen 1 is at such a low base in terms of Mhz "more than 10% higher performance" is frankly not even remotely unbelievable, i mean the best and highest clocked chip, the 1800X is clocked at 3.7Ghz all cores, if you bring that up to just 4.1Ghz its an 11% uplift in performance, i'm quit comfortable to say between 4 and 4.2Ghz would be Ryzen 2800X new base clock.

Intel are not the whole industry, some integraty may still exist outside of their offices, we have no reason to disbelieve Glofo or AMD on Ryzen.

pete910 · 10 Jan 2018 at 18:17

FFS the pair of you are a PITA