Well, hopefully it is an i7-8700k but no you did not. Enjoy your cpu and game on.
![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
Ryzen vs Coffee Lake
- Thread starter adobian
- Start date
Well, hopefully it is an i7-8700k but no you did not. Enjoy your cpu and game on.
Assuming you meant i7-8700k, definetly no. It is the best gaming CPU money can buy and a beast in worker things too. You have to pay big bucks but if you have the money it delivers.
If you meant i5-8600k, the answer is still no if you are a gamer only. For gaming it is almost on par with Kaby Lake i7, the king of last gen Intel. In situations where simply more threads rules the Kaby i7 is better but in situations where having more real cores is better than hyperthreading 8600k is even better than Kaby was. It is a great gaming CPU and OK as a worker CPU. Longevity is a question but that is always the case with i5. Hell, i5 7600K was outdated the moment it was released, 4 core 4 thread CPU's have no place in midrange gaming CPUs at 2017. Intel milked that cow way too long, so long that it is a dry husk now...
Best bang for the buck still belongs to Ryzen 5 1600. It is good for gaming and great for working. Ever since Coffee Lake was announced Ryzen 7 is in a little bit odd place right now. For a working CPU it is magnificient (better value than i7 Coffee Lake I think) but for gaming it is too expensive for the performance it gives because it is pretty much on par with Ryzen 5 hexcores right now. This is just an assumption but whenever gaming evolves to a point that the 8/16 threads of R7 becomes a clear boon over R5 6/12 the low clockseed of 1.gen Ryzens has probably become a problem by then and you are looking for upgrade anyway.
Hagrid
[H]F Junkie
- Joined
- Nov 23, 2006
- Messages
- 9,163
Basically if you are running 1080 60hz, Ryzen is fine. If you have a 120hz+ monitor and want fps to match(in all games), 8600k/8700k/7700k?/etc will do it.
Id still have the 7700k though, there is nothing out there it can't run and like previously mentioned it is still blazing fast per core performance and singlethreading, largely what a lot stuff still is etc.
Yeah, sorry, I meant i7. It was way overpriced n Finland in comparison with Rizen but I still went for it for the sake of "it is the best gaming CPU" I hear everywhere. And I do play on 100Hz G-Sync AOC AGon 35 Inch Ultrawide so it will be handy.
Yeah, everything is overpriced here in Finland. I am just about to build a Ryzen 5 rig. RAM prices are already expensive and the "Suomi-lisä" on top of them makes my heart bleed... But I think you made a good choice since you had the money for it (I don't so it is Ryzen 5 for me), that CPU should be future proof for very long time.
i7-8700k is better in almost everything, and when it loses to 1800X does by a tiny amount.
legcramp
[H]F Junkie
- Joined
- Aug 16, 2004
- Messages
- 12,405
Yes but it cost more.
It cost less.
legcramp
[H]F Junkie
- Joined
- Aug 16, 2004
- Messages
- 12,405
It cost more.
https://www.newegg.com/Product/Product.aspx?Item=N82E16819113430
It's only $349.99 right now at the awesome newegg breh, too bad you can't shop here huh? LOL
legcramp
[H]F Junkie
- Joined
- Aug 16, 2004
- Messages
- 12,405
whatever you say
legcramp
[H]F Junkie
- Joined
- Aug 16, 2004
- Messages
- 12,405
I venture to guess that in the end, both Intel and AMD are making products that about equal the other, in price and performance. Sure, there may be a sale here or deal there but all in all they are about the same price for what they are. In some areas, AMD holds the lead in performance and in other areas Intel holds the lead. If one only uses a PC to benchmark it, then get intel and your e-peen will be so happy. If you encode/transcode video while doing other stuff then AMD might be in your wheelhouse. If you game, either is a good choice since most games don't tax a CPU all that much at all. Pointless thread has turned more pointless.
That's what I said before...narrowed the gap ALOT more again like how it use to be..stuff like ThreadRipper and things. Just down to what's a better deal really, but clearly have shown they can make a Powerful Cpu.
IdiotInCharge
NVIDIA SHILL
- Joined
- Jun 13, 2003
- Messages
- 14,675
...and what do the maximum frametimes look like?
Because that's what you'll actually feel.
OrangeKhrush
[H]ard|Gawd
- Joined
- Dec 15, 2016
- Messages
- 1,673
20-30% in certain caveats, if running top of turbo or overclocked vs a stock ryzen in a load that is more affected by latency and clockspeed yeah maybe, same locked clocks its under 10%
Yeah, it's not really thattt significant. 7700k is and was still a few bucks cheaper than the 1800x, so kinda just came down to that.
8700k match or even beats 1800X in multithreaded benches. It is not difficult to see this implies CFL core has to be at least 33% faster than Zen core, not «20-30% in certain caveats».
Comparing clock for clock. Kabylake/CoffeeLake is 10--20% ahead of Zen with ancient software. The Stilt got Kabylake is 24% ahead of Zen with newest software
OrangeKhrush
[H]ard|Gawd
- Joined
- Dec 15, 2016
- Messages
- 1,673
8700k match or even beats 1800X in multithreaded benches. It is not difficult to see this implies CFL core has to be at least 33% faster than Zen core, not «20-30% in certain caveats».
Comparing clock for clock. Kabylake/CoffeeLake is 10--20% ahead of Zen with ancient software. The Stilt got Kabylake is 24% ahead of Zen with newest software
For general usage its like 12% which is about right, bearing in mind that these tests are normally stock and clock factor will play its part. This is why clock vs clock IPC cinebench test scores are hated, in general compute used by 90% of the mainstream marked, it is to close, enough for a nose bleed.
OrangeKhrush
[H]ard|Gawd
- Joined
- Dec 15, 2016
- Messages
- 1,673
It is very significant, clockspeed affects results more than cores do, it is like racing a Golf GTR against and an SLS AMG. IPC tests are always done clock vs clock and that has become a issue because reality is very different when you take away the 23-34% difference in clock speed base to turbo.
There is a reason why Floyd Mayweather didn't fight Mike Tyson even though pound for pound at mayweathers weight catogory he would beat just about anyone as it is easier going up than going down in weight class, similar thing, intel without the clockspeed it hides results behind it is not that far off.
If you take the 8700K CB ST score around 199 at 4.7ghz, then apply the test to Zen at the same clock the 1800x got 161~ at 4ghz that is at least 189 at 4.7ghz, just showing the effect of clockspeed. At 3ghz the 8700K scores 127 to Zens 120 even though the 1600 scores oddly enough 124.
Compare apples to apples or accept the reality that despite the handicap Zen is still rather punchy with Intels speed freaks.
Last edited:
That's why I said they do not understand man, still see a better response to a lot of things with Clocks. why the 4C/8T will be fine for some time. Realistically everything would have to be recoded and some implementation with more core/thread usage and that hardly ever happens even yet. Maybe sometime start seeing more stuff actually developed well to run like that. But largely as of yet the minimal scaling still seems to do perfectly fine.
I picked up the 1800x for $299 at Microcenter on Black Friday. I thought it was worthwhile at that price. I considered returning it for the 8700k, but I don't think it's worth the $100 more. The x370 Asrock Tiachi MB was $129 after the rebate, price drop, and $30 bundle savings.
For general usage its like 12% which is about right, bearing in mind that these tests are normally stock and clock factor will play its part. This is why clock vs clock IPC cinebench test scores are hated, in general compute used by 90% of the mainstream marked, it is to close, enough for a nose bleed.
That 12% is the IPC gap after eliminating the workloads where Zen does worse and then eliminating further the workloads with 256bit support. But there is no reason to eliminate those workloads to favor RyZen, It is so stupid as pretending that we cannot use recent versions of Blender or Handbrake only because they come with 256bit support.
When all the workloads are considered, Kabylake IPC is a 28.94% above Zen. We can round that to 29%.
So you pretend that the IPC gap in CineBench is 5% (199/189). But you are using wrong data. The 8700k does more than 200cb @4.7GHz. The 8700k does 220cb @ 5Ghz. The 1800X does 160cb @4.1GHz.
So the IPC gap in CineBench is much bigger than you pretend. A simple computations gives 12% IPC gap in CineBench, but to not get you confounded with clocks we have measurements at same clocks
IPC gap in Cinebench is 11%.
Finally no one is hating "IPC cinebench test scores". What people is explaining to you is that CineBench is a favorable benchmark for RyZen. Measurement made with CineBench do not represent measurements made with rest of hundred production workloads or games. What people is saying you is that you cannot take only CineBench scores and ignore everything else. To quote again from the review:
So stop mentioning only Cinebench scores and ignoring everything else, only because CineBench favor Zen.
The IPC gap in Cinebench is 11--12%, but the IPC gap in rest of workloads is 20--30%.
Last edited:
JustReason
razor1 is my Lover
- Joined
- Oct 31, 2015
- Messages
- 2,483
name calling will not be tolerated
Ok so what I get from this is BS, obfuscate, outright mislead, and more BS.That 12% is the IPC gap after eliminating the workloads where Zen does worse and then eliminating further the workloads with 256bit support. But there is no reason to eliminate those workloads to favor RyZen, It is so stupid as pretending that we cannot use recent versions of Blender or Handbrake only because they come with 256bit support.
When all the workloads are considered, Kabylake IPC is a 28.94% above Zen. We can round that to 29%.
So you pretend that the IPC gap in CineBench is 5% (199/189). But you are using wrong data. The 8700k does more than 200cb @4.7GHz. The 8700k does 220cb @ 5Ghz. The 1800X does 160cb @4.1GHz.
So the IPC gap in CineBench is much bigger than you pretend. A simple computations gives 12% IPC gap in CineBench, but to not get you confounded with clocks we have measurements at same clocks
IPC gap in Cinebench is 11%.
Finally no one is hating "IPC cinebench test scores". What people is explaining to you is that CineBench is a favorable benchmark for RyZen. Measurement made with CineBench do not represent measurements made with rest of hundred production workloads or games. What people is saying you is that you cannot take only CineBench scores and ignore everything else. To quote again from the review:
So stop mentioning only Cinebench scores and ignoring everything else, only because CineBench favor Zen.
The IPC gap in Cinebench is 11--12%, but the IPC gap in rest of workloads is 20--30%.
For everyone else please take note. The quoted post is what a shill post looks like. This is the end result of a poster that can not stand the fact that in REALITY, where a great deal of us live, the difference is minimal to non-existent. There are no links to graphs to attain time frame of results, where updates to software, bios, drivers or the like may have impacted results. Add to that the bickering over a % point or 2. And the belief that there is a 20-30%IPC gap in all other software other than Cinebench.
Here is a question for all you guys, does IPC really impact your daily lives? I am sure some it does, especially when their lively-hood depends on it. But how about us tinkerers and gamers? I doubt there is a difference of any measurable kind there. And before the idiocratic post about some 165Hz gaming, that only applies to about <0.05% of the market.
kirbyrj
Fully [H]
- Joined
- Feb 1, 2005
- Messages
- 30,710
I've said for a long time that the average user isn't going to notice a big difference, and a Ryzen 1600(x) at sub-$200 or a $230 Ryzen 7 is a phenomenal bang for the buck. I can't even get some of them to agree to the fact that AMD has decent performance in its price range (specifically the mid-range 1600 to 1700, not so much 1800x). That's kind of my litmus test. If you can't agree that AMD has a decent product at it's price range, you get put on ignore because you're an idiot. I've even had some of the 8700k supporters tell me that I was wrong for buying a 7820X because it doesn't have a higher clockspeed for single core gaming
.We get it. The Intel performs better in "many" tasks, specifically when you're pushing gaming to it's extreme levels and in single core workloads because of the high clockspeed. However, the difference isn't really noticeable for the average user. I wouldn't hesitate building a Ryzen system for someone on a budget.
jenosmaverick
Limp Gawd
- Joined
- May 24, 2017
- Messages
- 142
Uhm will Pinnacle Ridge with higher clockspeeds compared to the current Ryzens be decent with MMORPG games? MMO's are poorly optimize and single core heavy.
kirbyrj
Fully [H]
- Joined
- Feb 1, 2005
- Messages
- 30,710
Yes, they will be higher, but at this point nobody knows how much higher. If you need the single core performance, then look at a highly clocked Intel chip if you're buying right now. Maybe a 8600k?
jenosmaverick
Limp Gawd
- Joined
- May 24, 2017
- Messages
- 142
kirbyrj
Nah. My dan case will arrive in like 2 months right time for zen+ and the remainder of the coffee lakes. Hopefully zen+ do get better at gaming. At such tiny case R7 would be really sweet as an i7 would be really hot for it. If not might look into 8500 or 8400 even. Thanks
Nah. My dan case will arrive in like 2 months right time for zen+ and the remainder of the coffee lakes. Hopefully zen+ do get better at gaming. At such tiny case R7 would be really sweet as an i7 would be really hot for it. If not might look into 8500 or 8400 even. Thanks
OrangeKhrush
[H]ard|Gawd
- Joined
- Dec 15, 2016
- Messages
- 1,673
That 12% is the IPC gap after eliminating the workloads where Zen does worse and then eliminating further the workloads with 256bit support. But there is no reason to eliminate those workloads to favor RyZen, It is so stupid as pretending that we cannot use recent versions of Blender or Handbrake only because they come with 256bit support.
When all the workloads are considered, Kabylake IPC is a 28.94% above Zen. We can round that to 29%.
So you pretend that the IPC gap in CineBench is 5% (199/189). But you are using wrong data. The 8700k does more than 200cb @4.7GHz. The 8700k does 220cb @ 5Ghz. The 1800X does 160cb @4.1GHz.
So the IPC gap in CineBench is much bigger than you pretend. A simple computations gives 12% IPC gap in CineBench, but to not get you confounded with clocks we have measurements at same clocks
IPC gap in Cinebench is 11%.
Finally no one is hating "IPC cinebench test scores". What people is explaining to you is that CineBench is a favorable benchmark for RyZen. Measurement made with CineBench do not represent measurements made with rest of hundred production workloads or games. What people is saying you is that you cannot take only CineBench scores and ignore everything else. To quote again from the review:
So stop mentioning only Cinebench scores and ignoring everything else, only because CineBench favor Zen.
The IPC gap in Cinebench is 11--12%, but the IPC gap in rest of workloads is 20--30%.
Cinebench favours AMD? was this after they were busted for using codex in the bast that boosted intel scores artificially? Cinebench used to be the Intel bench of choice, now it is Tomb Raider lol.
You are looking at 10~ % either way IPC gap from Ryzen to Coffeelake at a lesser system cost, in parallel workloads Ryzen does well at lower clockspeeds and of course you can run the on Stock Coolers with the B350 chipset w ithout much sacrifice being made on features and clocking. If you get into Intel propriatory standards like AVX, it is clearer than day in Intels dosier that non Intel genuine parts only get baseline AVX support, but AVX is not a general computing standard, because nobody wants to be held to ransom by Intel corp. AVX is intel and those in bed with them, most end up getting burnt or get liquidated.
For day to day users and casual gamers a Ryzen is perfectly suited for anyone needing copious performance on modern standard hardware. But it is now rather pointless suggesting summit ridge mere months before pinnacle ridge.
IdiotInCharge
NVIDIA SHILL
- Joined
- Jun 13, 2003
- Messages
- 14,675
Decent performance for what?
The reason single-core performance is harped upon is because it is the one metric that an end-user is most likely to make use of. Gaming is a prime example, but not the only one.
An end-user may certainly notice the additional threaded resources per dollar with workloads other than gaming, of course, but in those cases the work (rendering? compiling? video editing?) simply takes a little more time, and if for that work time is money, then they're undercutting themselves by not going HEDT anyway. Below that range, gaming is the target.
There are so many wrong in this post and so little time...
(i) Some people is pretending since RyZen launch that the IPC gap between Kabylake and Zen is less than 10% or pretending that CineBench is all what matters to measure IPC, only Cinebench and nothing more. So IPC really impacts their lives; otherwise they wouldn't insist on IPC.
(ii) When certain people gets CB scores from their ass and post them here ("8700K CB ST score around 199", "1800x is at least 189"), you don't complain. Never! When I give benchmarks from mainstream reviews as PcPer, Arstechnica, HFR,... you always complain with the repetitive "There are no links to graphs to attain time frame of results, where updates to software, bios, drivers or the like may have impacted results" I like this double standard of yours.
(iii) Certain people pretended that the IPC gap in CB15 is only of 5%. Reviews given above show the IPC gap is about 12%. The difference between the 12% measured by reviews and the 5% someone else claimed is not "bickering over a % point or 2".
(iv) Lots of benchmarks from mainstream reviews show that the IPC gap is 20--30% in many workloads. Those reviews have been mentioned here. OrangeKrush also added a graph from a personal review made by The Stilt. The Stilt's review measured that Kabylake IPC is 29% ahead of Zen on average, when the full suite of benches is used to compute the average. So that 20--30% gap is not "a belief", as you claim, that gap is measured.
(v) I will stop here.
Last edited:
Learn to read. I didn't say that Cinebench favours AMD. I said "CineBench is a favorable benchmark for RyZen".
CineBench was broadly used because the IPC obtained agreed very well with the average obtained when testing IPC with different workloads. So instead running a dozen of different tests and getting the average, people preferred to just run CineBench and take that value as the average.
This doesn't work for RyZen chips. It continues working for other AMD chips as Piledriver, Jaguar, etc. No for RyZen.
Cinebench runs particularly well on RyZen. The performance measured using CineBench on RyZen don't correspond to the performance measured using dozens of other workloads. So you cannot take Cinenbench scores for RyZen and take that as the average performance because it is not. I will repeat this part from the Arstechnica review:
CineBench is a favorable case for RyZen, not the rule. The rule is that 8700k is faster.
Fine, you won the internet. Many cheers for the winner. May his prophetic prowess actually help society now instead of being wasted on a pointless thread.
Kabylake is too and the correct like a good 20-30%, cinebench just don't look good on Intel anymore 950-1100 scores and stuff.
OrangeKhrush
[H]ard|Gawd
- Joined
- Dec 15, 2016
- Messages
- 1,673
Arsrechnica the bible now? The same fools that ignore clockspeed factors more than core count? So a 30% higher clocked part should in most cases except in domains that actually scale full thread counts.
We have already done a full bench suite running the parts at 3ghz locked, we even simulated a real world consumer situation of workstations and used stock fans. Lets just say in a clock vs clock ipc showdown across the board it is extremely close.
OrangeKhrush
[H]ard|Gawd
- Joined
- Dec 15, 2016
- Messages
- 1,673
Pc gaming market is small the layman cannot be arsed about spec sheets so they go buy consoles, plug and play.
Mobility for corporate entities is by far the biggest segment followed by workstation.
Steam suggests modern socio economics, people using very old technology because they can't afford keeping up. This is why low spec games thrive or why console still sell to the "average joe"
Can I change the congratulations by a requirement?
The requirement is that here and thereafter people stop giving only CineBench scores for RyZen and ignoring any other benchmark where RyZen looks worse. To get he performance of a chip, any chip, one has to quote a collection of benchmarks, including games.
This is not about Arstechnica, neither about clocks. Other reviews find the same findings than Arstechnica. It is not about clocks because the clocks are the same when you run Cinebench than when you run Blender or Handbrake or 7-zip... The 1800X is a 3.6/4.0/4.1GHz chip in all cases.