Broadwell-E vs Ryzen in games (PurePC)

You're missing his point entirely. With me being in the situation that I've got a 2600k predominantly game and have a serious itch for a real CPU upgrade, the whole Rysen lineup offers me nothing new as far as performance goes from what i've already had from 2011. With that said neither are Intel at the moment, and that is a bad thing. Looking forward to new CPU releases.

Huh? Well, maybe you do something different with your computer than I do, but double the cores & threads, plus an IPC bump? That does nothing for you? I upgraded my 2600k system and see close to *three times* the encoding and rendering performance with my Ryzen 1700X.

I guess if you're just gaming, there's no need for that. But then, if you're just gaming, your 2600k is just fine, if you slap in a fat GPU. Note: the 6c and 8c Ryzens will still be faster in games than the 2600k. So will the 4c 1500X, though not by enough to be worth it, IMHO.

In other words, your complaint would be equally valid with Intel. 7700k is ~20% better IPC than Sandy Bridge (according to Kyle's own tests), and OCs maybe 400-500MHz more. So, at best, you're looking at a 30-40% CPU performance bump with a 7700k over a 2600k. I was in the same boat as you... not really worth a complete system rebuild for 30-40% CPU gains. But if you have anything which can use more cores, Ryzen starts to look real nice. Double to triple the encoding & rendering performance compared to my 2600k? Oh yeah, I'll take that all day long, man. All day.
 
Huh? Well, maybe you do something different with your computer than I do, but double the cores & threads, plus an IPC bump? That does nothing for you? I upgraded my 2600k system and see close to *three times* the encoding and rendering performance with my Ryzen 1700X.

I guess if you're just gaming, there's no need for that. But then, if you're just gaming, your 2600k is just fine, if you slap in a fat GPU. Note: the 6c and 8c Ryzens will still be faster in games than the 2600k. So will the 4c 1500X, though not by enough to be worth it, IMHO.

In other words, your complaint would be equally valid with Intel. 7700k is ~20% better IPC than Sandy Bridge (according to Kyle's own tests), and OCs maybe 400-500MHz more. So, at best, you're looking at a 30-40% CPU performance bump with a 7700k over a 2600k. I was in the same boat as you... not really worth a complete system rebuild for 30-40% CPU gains. But if you have anything which can use more cores, Ryzen starts to look real nice. Double to triple the encoding & rendering performance compared to my 2600k? Oh yeah, I'll take that all day long, man. All day.


30-40% is substantial, the problem with intel is that to get equally good you have to buy substantially more expensive boards. If you plan on overclocking only the top end boards get you close or to 5Ghz.
 
30-40% is substantial, the problem with intel is that to get equally good you have to buy substantially more expensive boards. If you plan on overclocking only the top end boards get you close or to 5Ghz.

30-40% is middling. If it were a drop-in replacement for my old 2600k, I'd have considered it. Requiring new mobo & RAM in addition to the CPU... it starts to look kind of "meh." If Ryzen hadn't come along, I probably would have gone with a 7700k, but that'd be more because my old Sandy Bridge system was just unstable after 6 years of abuse. It was on its way out, and needed replacement anyway.
 

Thanks! I like HFR reviews because they always give average performance figures. Their measurements for the R5 1600 agree very well with predictions made before launch by 3DCenter: 6C Ryzen @3.3GHz ~ 4C KBL @4.2GHz

getgraphimg.php



AMD-Ryzen-geschaetzte-Workstation-Performance.png
 
Thanks! I like HFR reviews because they always give average performance figures. Their measurements for the R5 1600 agree very well with predictions made before launch by 3DCenter: 6C Ryzen @3.3GHz ~ 4C KBL @4.2GHz

So according to this review AMD is at or below Sandy Bridge core performance in games (4C/8T 2600K ~ 4C/8T R5 1500X).
 
So according to this review AMD is at or below Sandy Bridge core performance in games (4C/8T 2600K ~ 4C/8T R5 1500X).

Don't know where you're getting "at or below" when the 1500X is actually 3.7% faster than the 2600k in games, in the very graph you're citing.
 
Don't know where you're getting "at or below" when the 1500X is actually 3.7% faster than the 2600k in games, in the very graph you're citing.

Everyone at [H] generally factors in overclocking since it's in the name of the forum name itself. Ryzen gets wrecked which is kinda disappointing to me. Basically a 3770k can beat a 1800x OC vs OC and I sold that processor years ago...

Still an option for friends rigs I suppose.
 
Everyone at [H] generally factors in overclocking since it's in the name of the forum name itself. Ryzen gets wrecked which is kinda disappointing to me. Basically a 3770k can beat a 1800x OC vs OC and I sold that processor years ago...

Still an option for friends rigs I suppose.

Yep I agree, the kids Haswell is running at 4.8 and there's literally zero reason to change it because all he does is game and homework. Nothing "productive" unless you count the boxes of tissue going missing from around the house. RB6S and modded Skyrim isn't going to benefit from more lower speed threads, it'd be a step back.
 
Yep I agree, the kids Haswell is running at 4.8 and there's literally zero reason to change it because all he does is game and homework. Nothing "productive" unless you count the boxes of tissue going missing from around the house. RB6S and modded Skyrim isn't going to benefit from more lower speed threads, it'd be a step back.

If I were running a Haswell @ 4.8, and just gaming and bulsh*tting with it, I wouldn't upgrade. Not to a Ryzen, or to a Kaby Lake. Just not worth it IMHO.
 
1500X: 118.7
2600k: 115.7

The gap is 2.6%. And that is at stock clocks, the i7 OC better.

My math sucks today. Anyway, I already conceded that point. In fact, I'll go one further, and suggest maybe I should've gone ahead and spent the $1,050 on a Broadwell-E instead of a Ryzen. But it was difficult to pull the trigger on $650 higher price for such a small jump.
 
You know... I think y'all just legit changed my mind on this. Maybe Ryzen really did get wrecked, except in very specific workloads. And I'm just kind of blind to it, because I do a lot of those specific things.

Wish I had more time for games. But that'll probably never happen.
 
Last edited:
You know... I think y'all just legit changed my mind on this. Maybe Ryzen really did get wrecked, except in very specific workloads. And I'm just kind of blind to it, because I do a lot of those specific things.

Wish I had more time for games. But that'll probably never happen.

In terms of picking for an upgrade for gaming sure it got "wrecked" because it's not an upgrade. A lot of Intel's recent releases aren't upgrades either no matter how the coin is spun.

What they released was very specific in an ordered way to test their fabrication and design work... chips without graphics combining two bundles of cores across their fabric. Such a guinea pig chip. The bonus was, the cores don't suck and you get lots of them comparatively cheap, you just don't get OC or high clockspeed with them and you have to adopt the growing pains of waiting for software that is Ryzen optimized. Neither the Ryzen NOR the "prosumer" line is going to interest me personally, at all, ever.

I want Naples, and I want strong HTPC APU's. At least having seen these cores in action on Ryzen I'm moderately confident those products will be very decent and price competitive.
 
Nah, I'm starting to think the Intel fanboys have it right. Or at least mostly right. When it comes right down to it, the chip at the top of most of the charts is an Intel chip. Sure, Ryzen beats the top-of-the-line Mainstream Intel chip in threaded situations. And sure, Ryzen beats the value of the HEDT Intel chip if money is an object (otherwise I wouldn't have bought one). But to achieve that AMD had to make a lot of sacrifices. I think Sweeper and juanrga successfully defended their position: Ryzen is just an 8 core Sandy Bridge on a modern platform.

And so Intel wins the race, in the end.

But hey, I got a cheap 8 core chip out of the deal, so it's not bad for me.
 
But that's what the entire Ryzen line is... just another chip competing on price. Nothing to really differentiate it, yet. That's what I was talking about back a while ago

Yeah, you're right on that.

I posted some time ago about replacing my dying 2600k rig... 7700k, Ryzen, or 6900k. Went back and forth. Settled on Ryzen eventually. I did right by avoiding the 7700k since threaded performance wasn't there and I wanted that. But I think if I were to do it again, I'd fork over the cash and buy a 6900k instead. That way I wouldn't have to make any compromises on the rig. But I'm reasonably pleased with the Ryzen rig anyway, and maybe the AM4 socket will see a good size bump down the road.
 
Yeah, you're right on that.

I posted some time ago about replacing my dying 2600k rig... 7700k, Ryzen, or 6900k. Went back and forth. Settled on Ryzen eventually. I did right by avoiding the 7700k since threaded performance wasn't there and I wanted that. But I think if I were to do it again, I'd fork over the cash and buy a 6900k instead. That way I wouldn't have to make any compromises on the rig. But I'm reasonably pleased with the Ryzen rig anyway, and maybe the AM4 socket will see a good size bump down the road.

Now you have me confused, on a purely productivity level with Ryzen vs 6900K, what did you give up? I asked you before what you'd pick if they were the same price, but they're not the same price. To the point where you can buy a top shelf GPU with the difference. I mean that's the one area where it makes a ton of sense at that price.
 
Now you have me confused, on a purely productivity level with Ryzen vs 6900K, what did you give up? I asked you before what you'd pick if they were the same price, but they're not the same price. To the point where you can buy a top shelf GPU with the difference. I mean that's the one area where it makes a ton of sense at that price.

Nothing. Well... maybe a small amount, since the 6900k OCs a little better. Call it, say, 10% performance difference after OC. But I do still game... sometimes. So I gave up some FPS there too. And I already have a top shelf GPU. Or, second tier, anyway, since the Titan Xp (what a stupid name, btw) came out to edge out the 1080 Ti.

But the last little bit of performance is always expensive. You gotta pay to play. That's the Intel way ;).
 
6950k OCs at least 22%. 1800x really isn't in the same ballpark. Plus you get great gaming and application performance.

Is it worth it for me? No... but a lot of people that can save 40%ish time in productivity apps don't care about a grand.

Ryzen also has a memory limitation. Ryzen is in a weird position. Now a 16C/32T Ryzen could be very interesting especially if it's priced around the 6950k, but I doubt it will be.
 
6950k OCs at least 22%. 1800x really isn't in the same ballpark. Plus you get great gaming and application performance.

Is it worth it for me? No... but a lot of people that can save 40%ish time in productivity apps don't care about a grand.

Ryzen also has a memory limitation. Ryzen is in a weird position. Now a 16C/32T Ryzen could be very interesting especially if it's priced around the 6950k, but I doubt it will be.

It won't be, I'm sure.

Oh well. Another dud, and another several years of ho-hum ticks, tocks, and 5% bumps await us.

I miss the old days, sometimes.
 
Last edited:
6950k OCs at least 22%. 1800x really isn't in the same ballpark. Plus you get great gaming and application performance.

Is it worth it for me? No... but a lot of people that can save 40%ish time in productivity apps don't care about a grand.

Ryzen also has a memory limitation. Ryzen is in a weird position. Now a 16C/32T Ryzen could be very interesting especially if it's priced around the 6950k, but I doubt it will be.

With SKL-X around 2½ months away we all know what the real competition is. From a server perspective, I can tell SKL-EP is bigger than what Nehalem was.
 
1500X: 118.7
2600k: 115.7

The gap is 2.6%. And that is at stock clocks, the i7 OC better.

Actually, I just realized that you are bad at math too. The gap is 3% given that the performance is normalized to a 2500k as 100%.

Not that this really means anything. But if you're going to pick a nit, I'll do the same ;).
 
With SKL-X around 2½ months away we all know what the real competition is. From a server perspective, I can tell SKL-EP is bigger than what Nehalem was.

If 16c Ryzen fails, AMD will be out of business. They bet the farm on the Zen architecture.

Then we can all enjoy high Intel pricing forever.
 
6950k OCs at least 22%. 1800x really isn't in the same ballpark. Plus you get great gaming and application performance.

Is it worth it for me? No... but a lot of people that can save 40%ish time in productivity apps don't care about a grand.

Ryzen also has a memory limitation. Ryzen is in a weird position. Now a 16C/32T Ryzen could be very interesting especially if it's priced around the 6950k, but I doubt it will be.

6950k was too much. But I could have swung a 6900k @ $1k. And I should have done that. I should turn in my enthusiast card for buying a slow CPU. But what can I say? I'm poor. And that's always a good argument for buying AMD.

But I already have this Ryzen rig, and it is more than twice as fast as my old 2600k in productivity sh*t, and it OCs to 4GHz, and it was cheap, so I'll just stick with it.
 
If 16c Ryzen fails, AMD will be out of business. They bet the farm on the Zen architecture.

Then we can all enjoy high Intel pricing forever.

Are you talking about the leaked prosumer line and not Naples? Still won't be much good for me. Naples though.... that will be, let's see the price. I'll take 48 or 64 threads with 256 Gigs of RAM on an eATX board please.

And really AMD, if successful with this architecture, is going to make their bread and butter from the server and APU line. These Ryzen chips are only good for discrete graphics users and that isn't all that much of the market overall.

From an Anantech article "In the second quarter of 2016, the industry sold between 62.4 million and 64.3 million PCs, according to IDC and Gartner respectively. Meanwhile, discrete GPUs were used in 27.78% of PCs, claims JPR."

We're seeing that Ryzen and Ryzen prosumer will only be eligible for (rounding up) to 30% of the market. That's great etc... blah blah. But that other seventy per cent of the market needs their IGPs in a bad way and every review of an AMD APU says the same thing "The graphics while theoretically fast enough for blah blah are handcuffed by the lackluster CPU cores and heavy power usage". This is the point at which Zen is making sense... four of these cores and a bunch of polaris type CU's on APU's in the first round. Zen+ & Vega CU's in the second round....

I mean from what we've seen, let's say we have an APU with Four Zen cores @2.2-3.2Ghz + a half downclocked polaris all on DDR4 , it's not going to be a bad APU halo product. I really have hope for AMD but it's not Ryzen that matters all that much (it only matters 30%), Zen matters and it's not bad and it's not out of the power/performance footprint like Dozer+ cores.
 
Are you talking about the leaked prosumer line and not Naples? Still won't be much good for me. Naples though.... that will be, let's see the price. I'll take 48 or 64 threads with 256 Gigs of RAM on an eATX board please.

And really AMD, if successful with this architecture, is going to make their bread and butter from the server and APU line. These Ryzen chips are only good for discrete graphics users and that isn't all that much of the market overall.

From an Anantech article "In the second quarter of 2016, the industry sold between 62.4 million and 64.3 million PCs, according to IDC and Gartner respectively. Meanwhile, discrete GPUs were used in 27.78% of PCs, claims JPR."

We're seeing that Ryzen and Ryzen prosumer will only be eligible for (rounding up) to 30% of the market. That's great etc... blah blah. But that other seventy per cent of the market needs their IGPs in a bad way and every review of an AMD APU says the same thing "The graphics while theoretically fast enough for blah blah are handcuffed by the lackluster CPU cores and heavy power usage". This is the point at which Zen is making sense... four of these cores and a bunch of polaris type CU's on APU's in the first round. Zen+ & Vega CU's in the second round....

I mean from what we've seen, let's say we have an APU with Four Zen cores @2.2-3.2Ghz + a half downclocked polaris all on DDR4 , it's not going to be a bad APU halo product. I rally have hope for AMD but it's not Ryzen that matters all that much (it only matters 30%), Zen matters and it's not bad and it's not out of the power/performance footprint like Dozer+ cores.

Beating Bulldozer is no great achievement. AMD should never have come out with that pile of crap. AMD was good once, man. They really were. Even the Phenoms... okay, they were losing to the first-gen Nehalems, but not by *that* much. And AMD had been switching places with Intel between '99 and '07 all the time. They threw away everything with Bulldozer, and I still have to wonder why.

Zen still kind of sucks, I guess. The Intel fanboys seem closer to the truth on this. It's just that Zen sucks a lot less than Bulldozer, and might actually make sense for some people with some workloads. So compared to old AMD, new AMD looks great. But relative to Intel, they still suck. But to be fair to AMD, they had to design this thing on a shoestring after they sold off their own fabs to stay afloat. I mean, it's pretty damned good when you take all that into consideration. Imagine if they released something decent back in 2011. Maybe not an Intel killer, but something competitive that sold better. Maybe we'd be in a better spot today.

I don't know. I'm pretty disillusioned with the CPU market in general. It's felt stagnant and sh*tty ever since Ivy Bridge. Sandy Bridge was the last really exciting thing to come out, IMHO. I mean I see folks in this thread talking about an awesome Skylake-X, and I'm just thinking how hard it is to get excited about a minor IPC and clock bump (zOMG our new design is 4.0 GHZ instead of 3.6 GHz it's teh awesomeZ). Used to be, we got excited for a 50% clock jump, or a huge IPC boost, or those days when dual cores and quad cores were new, and it was great.

I mean at least AMD fanboys have something to get excited about. AMD wasn't bullsh*tting about a 52% IPC improvement over Crapdozer.

Nvidia keeps chugging with GPUs, though. So there's that.
 
Last edited:
About the only workload I do that really takes advantage of 8 cores is Handbrake encodes, and the chip I did that on before the 5960X @ 4.6GHz was a 3770K @ 4.6GHz. There's no doubt in my mind that any 8 core Ryzen would mop the floor in Handbrake vs that 3770K. So I still think Ryzen is a great chip for somebody still chugging along with HB on an old quad core Intel, and I think Ryzen was a huge step up for AMD. But I've also said from the get go that knocking off the TOTL Intel HEDT chips was going to be a stretch - look back in the AMD CPU forum (where the Ryzen discussion generally takes place) at all the hype laden headlines of Zen crushing this and beating that.

And while Ryzen is a good showing, and can compete on a clock to clock basis, it didn't really crush the TOTL Intel SKUs. If it would overclock, then we'd be having a whole different conversation. But that's where Intel had the past experience to build on. You complain about little 5% tics and tocks, but that's product development. That's why Intel has been running quad channel memory on 8c/16t chips that will also run a 50% OC now for a couple of years. I expect AMD will learn a few things from Ryzen and show marked improvement in their next generation flagship CPU, just like Intel had to claw their way back from former AMD superiority.
 
About the only workload I do that really takes advantage of 8 cores is Handbrake encodes, and the chip I did that on before the 5960X @ 4.6GHz was a 3770K @ 4.6GHz. There's no doubt in my mind that any 8 core Ryzen would mop the floor in Handbrake vs that 3770K. So I still think Ryzen is a great chip for somebody still chugging along with HB on an old quad core Intel, and I think Ryzen was a huge step up for AMD. But I've also said from the get go that knocking off the TOTL Intel HEDT chips was going to be a stretch - look back in the AMD CPU forum (where the Ryzen discussion generally takes place) at all the hype laden headlines of Zen crushing this and beating that.

And while Ryzen is a good showing, and can compete on a clock to clock basis, it didn't really crush the TOTL Intel SKUs. If it would overclock, then we'd be having a whole different conversation. But that's where Intel had the past experience to build on. You complain about little 5% tics and tocks, but that's product development. That's why Intel has been running quad channel memory on 8c/16t chips that will also run a 50% OC now for a couple of years. I expect AMD will learn a few things from Ryzen and show marked improvement in their next generation flagship CPU, just like Intel had to claw their way back from former AMD superiority.

AMD lost, that's for sure. And they should have had quad channel RAM for Ryzen. It clearly likes fast RAM, and lots of RAM bandwidth. As for OC... anyone who expected these chips to be great overclockers was fooling themselves. AMD could not afford to leave any performance on the table, as they were playing catch up.

AMD still sucks. They are still slow. They are still a distant #2. But Intel can at least see them in the rearview mirror now, and that's a huge improvement for AMD.

The only argument for AMD is still price. They'll sell you more cores for less money, and unlike Bulldozer cores, they aren't total sh*t. But it's still basically an 8 core Sandy Bridge, when all is said and done. That made some sense for me, because I'm poor. But for most, I've changed my mind on this sh*t: buy a 6900k or a 7700k, depending on your needs. Don't waste your time with Ryzen *if* you have the money to spend. Only look at Ryzen if you're trying to save cash and could use the extra cores for something.
 
30-40% is middling. If it were a drop-in replacement for my old 2600k, I'd have considered it. Requiring new mobo & RAM in addition to the CPU... it starts to look kind of "meh." If Ryzen hadn't come along, I probably would have gone with a 7700k, but that'd be more because my old Sandy Bridge system was just unstable after 6 years of abuse. It was on its way out, and needed replacement anyway.
I fully agree with you. The CPU cost is okay, but the platform cost is a killer, especially with RAM factored in. No matter which one I look at, Ryzen or Intel, the cost to upgrade is not justified in any practical terms.

What I am doing at the moment is optimizing my current storage solution. After that, I may have enough funds left over to start considering a platform upgrade, but that will come later in the year.
 
I fully agree with you. The CPU cost is okay, but the platform cost is a killer, especially with RAM factored in. No matter which one I look at, Ryzen or Intel, the cost to upgrade is not justified in any practical terms.

What I am doing at the moment is optimizing my current storage solution. After that, I may have enough funds left over to start considering a platform upgrade, but that will come later in the year.

That's what I was getting at about lack of CPU excitement since Sandy Bridge. Yeah, as Dullard said, the 5% ticks and tocks were improvements, but they were nothing like the blistering pace we had from the 90s up to SB. There really isn't much of a compelling reason to upgrade 6 year old machines, unless the machine is failing, or unless you've got sh*t that can use more cores. Back in the mid 2000s, a 6 year old machine was ancient. Today it's like... middle of the pack. Pretty damned OK.

I don't care what anyone says about that situation, or what excuses they make for both Intel and AMD... it sucks for enthusiasts.
 
I don't care what anyone says about that situation, or what excuses they make for both Intel and AMD... it sucks for enthusiasts.

I agree with you , it does suck for enthusiasts.

So when the trolls are out in full force wishing the death of AMD and just leaving us with Intel and Nvidia, making it even worse, I just realize how incredibly stupid and young they are. Some are obviously paid trolls, some are pretenders, all are wrong. Calling a spade a spade and saying that a product doesn't live up to what you need or desire, that's great feedback and you gave some of that. Wishing there was no competition at all when we have so very little of it now is just plain stupid. I ignore those folks heavily.

I'm still fairly confident this Zen step will save AMD's life, but I never thought they'd be battling for the performance crown and low and behold they aren't . What they offer for the money is arguably reasonable though on some of the stack.

But yeah, those enthusiast glory days when I was building three stage cascades are long over with. It is what it is. All hail the next 3%.
 
Calling a spade a spade and saying that a product doesn't live up to what you need or desire, that's great feedback and you gave some of that...
I'm still fairly confident this Zen step will save AMD's life, but I never thought they'd be battling for the performance crown and low and behold they aren't . What they offer for the money is arguably reasonable though on some of the stack.

Ryzen is good for me. Not the best. But good. Really, I shouldn't be complaining, because it really is double and sometimes triple the performance of my old 2600k in rendering/encoding and other multi-threaded workloads. And that's where the most time got wasted. It's just that with Ryzen, you leave some performance on the table, whereas with Intel, you get all of it. Less second-guessing if you buy the best, I suppose.

But again, as you said, it's only the pricing that's going to save it, and by extension, AMD itself. They got close enough to make the proposition interesting for a lot of folks, so maybe they'll do okay. And really, I hope they do. Nvidia has done pretty well, even though they have market dominance right now. But Intel definitely got pretty lazy. And if AMD was gone tomorrow, I've no doubt that Nvidia would get just as lazy too.
 
Well, ryzen was a huge upgrade over the locked clock broadwell X99 i was using. Except for select VR games i was plenty happy with a stock clocked fx-8320e. While Ive owned the chip, ive only run a few benchmarks and gamed maybe 6-8 hours in the last 6 weeks -- i am looking forward to encoding 4k 100mbps video where ill see some real benefits from this platform
 
And I still don't think Ryzen sucks (although it is baffling why this conversation is still in the Intel subforum), it just had an unreachable level of hype leading up to its launch. I've seen enough productivity numbers to tell me it's damn stout for something like video encoding, and will open up huge leaps of ability to do that sort of work for basically the same price people were paying for half the cores. Obviously not enough to make me ditch a very solid 5960X rig for, but I guarantee you if I was poking along doing encodes on a 4 core SB or IB and wanted to cut my time in less than half by building a new rig, then Ryzen would be an easy choice.

So don't get all down in the dumps, I'm still very impressed with the numbers it's actually putting up - even if they do fall short of what the pre-launch hype would have us believe.
 
And I still don't think Ryzen sucks (although it is baffling why this conversation is still in the Intel subforum), it just had an unreachable level of hype leading up to its launch. I've seen enough productivity numbers to tell me it's damn stout for something like video encoding, and will open up huge leaps of ability to do that sort of work for basically the same price people were paying for half the cores. Obviously not enough to make me ditch a very solid 5960X rig for, but I guarantee you if I was poking along doing encodes on a 4 core SB or IB and wanted to cut my time in less than half by building a new rig, then Ryzen would be an easy choice.

But a 6900k or 6950x would still be better, in the end. Just a lot more expensive. Ryzen has a place, but it's place is in the mid-range, competing primarily on price, not on performance. It has just enough on the latter to make price competition possible.

So don't get all down in the dumps, I'm still very impressed with the numbers it's actually putting up - even if they do fall short of what the pre-launch hype would have us believe.

I'm more down in the dumps about how lame the CPU market has been in general. Every time I would pop out of the woodwork, every couple of years, to see if my old 2600k was worth upgrading, the answer was always "nope, not really." This time it was "fine, fine, I'll do it, but mostly because it's about to f*cking die after 6 years of abuse." Double the cores is nice, don't get me wrong. And each one is as fast as my old Sandy Bridge.

But damnit, I expected more out of the market after 6 f*cking years. Choices were *another* Intel quad that's like only 30% faster, an 8 core AMD that's no faster at all per thread (but you do get 8 cores), and being bent over for $1000+ for a properly quick Intel 8 core. Crappy choices man. I said in my original thread... no matter what I chose, I was going to have a lot of regrets on the purchase.
 
I mean, maybe y'all don't see it like I do. I stepped out of the enthusiast world for many years. Got married, had kids... whatever. Stopped gaming much. Didn't have time for OC adventures and stupid stunts like I used to do. I'd do periodic quick checks to see if time for an upgrade, but didn't stay around long. So I come back, because it's time for a new build, and I've got years of catching up to do...

And I'm all "what? That's IT? That's 6 years worth of R&D? WTF happened?" Aside from GPUs (they did great), advancement was more or less at a standstill. I ran my old 2600k at 4.2, once upon a time, though I could have pushed it a bit further, maybe. There's Kaby Lake, running at 4.2 stock, with maybe a 20% IPC improvement. Granted, some folks are hitting 5GHz (but some folks were doing 4.5 on Sandy). But still... that's IT? Broadwell is what, 14% better IPC than Sandy? Clocks not a lot higher either, though you at least get 8 and 10 core models. Now, AMD was complete garbage when I vanished. I come back and they are much better, but still nowhere near their peak in the old days. Lame, lame, and more lame.

In May of 1999, Intel released the 550MHz Katmai P3
In May of 2005, AMD released the 2400MHz Athlon 64 X2 dual core (4800+)

Now THAT was some serious movement in the market in 6 years! And 6 years after THAT, we got quad core Sandy Bridge @ 3.5GHz. Again! Lots of movement!


Now, I'm reading that maybe we're running up on the end of silicon. That 7nm is as far as it's going to go, and IPC is nearing a wall (because of branch prediction, etc...). So maybe we can't go much further. But even if so, it still sucks.

Okay, I'm done being negative for a while. At least my 1080 Ti trounces the old 7970 thoroughly. That's the bright spot in all this.
 
Last edited:
Nah, I'm starting to think the Intel fanboys have it right. Or at least mostly right. When it comes right down to it, the chip at the top of most of the charts is an Intel chip. Sure, Ryzen beats the top-of-the-line Mainstream Intel chip in threaded situations. And sure, Ryzen beats the value of the HEDT Intel chip if money is an object (otherwise I wouldn't have bought one). But to achieve that AMD had to make a lot of sacrifices. I think Sweeper and juanrga successfully defended their position: Ryzen is just an 8 core Sandy Bridge on a modern platform.

And so Intel wins the race, in the end.

But hey, I got a cheap 8 core chip out of the deal, so it's not bad for me.

To be fair, most of Intel's lineup is just a Sandy Bridge rehash. Apart from increasing base clocks (which means little to the overclocker, given how crazy SB OC'd), Intel hasn't done much either. Ryzen got AMD back to a relevant position. The real question here is if they can scale the Zen architecture upwards at a reasonable rate from here. If they're forced to do an Intel like "5% more IPC, 5% more clock speed" level of slow improvements, they're not gonna be in great shape. If Zen has much larger gains to be got faster, things could be better. Guess we'll see.

I think the supposed mainstream 6 core Coffee Lake is probably the first CPU to get semi excited about lately, at least for the average enthusiast.
 
My bet is that the next Ryzen - whatever it's called - will be the one to sprout wood over.
 
Back
Top