Who's planning to buy Zen?

[Magic Hate Ball]

In my case, I have the heart of an enthusiast, but not the wallet. I don't upgrade until I start to really feel like my rig is holding me back. It's usually around 4-5 years, with a GPU upgrade halfway through.

Previously, you would see HUGE gains from that upgrade. If we go back to all my rigs (the first couple I was just a kid, so more my parents rigs), I went from a Pentium 133, to a P3 800, to a P4 2.8Ghz, to a Core 2 Duo (2.4ish GHz iirc), to a 2500k that I run modestly at 4GHz. During each one of those swaps, the rig was getting barely tolerable by the time the next one was purchased. Right now I know my 2500k is long in the tooth, and there's definitely features I'm missing, but there's nothing unusable about it atm. It plays every game I throw at it just fine and I don't do anything too CPU intensive beyond that.

I'll probably build a new rig soon with either a 7700k or Zen, we'll see how that turns out. Not because I really feel like I need to yet, but because my Dad gets my hand-me-downs and that C2D system is PAINFUL.

This exactly. In the past I only moved for significant performance increases (K6-2 500mhz >> T-Bird Athlon 1.33ghz >> Athlon64 3200+ >> Core 2 Duo E8400 >> i7-875k) but right now I'm sitting on an i7-4770K I got super cheap ($125 for CPU+Mobo+16gb of RAM) that I would not have upgraded to without the deal as the performance bump just wasn't that huge.

I was going to wait for Zen but this came up and I didn't want to pass it up. I may resell it and buy Zen as I can probably sell it for significantly more than I bought it for still as I really want to support AMD's turnaround to keep competition in the market buoyed.

 

[maxius]

CC cleared out and waiting

 

[Hellstoaster]

Even if you have the need to build a new system you would not consider Ryzen if it is close to Intel and priced attractively?

Only a fool wouldn't consider all options when upgrading their system. You say "close to Intel" and "priced attractively." From whats been shown so far, if you have an i7 from the last 4 years Ryzen isn't going to get you anything and AMD hasn't been saying much in terms of pricing. AMD is limping along financially, just how cheap do you think they will price their new flagship chips? They are desperate for cash and they aren't going to cut their own throat to spite Intel. If they have a chip that is competitive with Intel's flagship SKU's then it will be priced similar to Intel's chips.

People running Sandy Bridge, Ivy Bridge, or AMD's own FX chips should be looking at Ryzen. Anyone with a Haswell or better probably won't get much, if anything, out of it.

If pricing and performance are comparable of course AMD's chips will be up for consideration but I don't expect any miracles on the performance or pricing front.

 

[Brackle]

Only a fool wouldn't consider all options when upgrading their system. You say "close to Intel" and "priced attractively." From whats been shown so far, if you have an i7 from the last 4 years Ryzen isn't going to get you anything and AMD hasn't been saying much in terms of pricing. AMD is limping along financially, just how cheap do you think they will price their new flagship chips? They are desperate for cash and they aren't going to cut their own throat to spite Intel. If they have a chip that is competitive with Intel's flagship SKU's then it will be priced similar to Intel's chips.

People running Sandy Bridge, Ivy Bridge, or AMD's own FX chips should be looking at Ryzen. Anyone with a Haswell or better probably won't get much, if anything, out of it.

If pricing and performance are comparable of course AMD's chips will be up for consideration but I don't expect any miracles on the performance or pricing front.

tbh after seeing the 7700k vs 2600k @ 4.5ghz. Anyone with a Sandy/Ivy bridge should be fin.e I would say anyone with an FX should upgrade.

Hell even my Xeon 5670 is still holdings its own!

I do plan to upgrade to ryzen IF (BIG IF) it lives up to the hype. How could you not want a 8c/16t CPU you can overclock. I'm a hardware nerd, and this has peaked my interest....but ONLY if the chip isn't a flop. That is why I plan to wait for the benchmarks and Kyle's review.

I would rather spend $500 on an 8c/16t CPU then $1100.....

 

[Hellstoaster]

tbh after seeing the 7700k vs 2600k @ 4.5ghz. Anyone with a Sandy/Ivy bridge should be fin.e I would say anyone with an FX should upgrade.

Hell even my Xeon 5670 is still holdings its own!

I do plan to upgrade to ryzen IF (BIG IF) it lives up to the hype. How could you not want a 8c/16t CPU you can overclock. I'm a hardware nerd, and this has peaked my interest....but ONLY if the chip isn't a flop. That is why I plan to wait for the benchmarks and Kyle's review.

I would rather spend $500 on an 8c/16t CPU then $1100.....

I figured people with Sandy/Ivy setups would consider it just for the new platform with all the new bells and whistles like m.2 and updated USB and such. Ryzen will probably offer some performance increases on those chips but also provide access to the new toys.

8c/16t sounds good, yes, but it's also completely useless in gaming and daily tasks. If someone was updating their workstation and used applications that could make use of that many threads it would be a tempting upgrade. Also I genuinely don't believe they will get close to 6900k performance and only charge $500. As desperate as they are for money they will price it closer to the 6900k. Maybe not as outrageous as $1100 like Intel did, but it will be higher than $500. The only reason they would price it at $500 is if it was lacking in some capacity and they were begging people to buy it.

 

[Brackle]

I figured people with Sandy/Ivy setups would consider it just for the new platform with all the new bells and whistles like m.2 and updated USB and such. Ryzen will probably offer some performance increases on those chips but also provide access to the new toys.

8c/16t sounds good, yes, but it's also completely useless in gaming and daily tasks. If someone was updating their workstation and used applications that could make use of that many threads it would be a tempting upgrade. Also I genuinely don't believe they will get close to 6900k performance and only charge $500. As desperate as they are for money they will price it closer to the 6900k. Maybe not as outrageous as $1100 like Intel did, but it will be higher than $500. The only reason they would price it at $500 is if it was lacking in some capacity and they were begging people to buy it.

So basically you don't think games take advantage of more cores/thread (new games do), and you do not think it will come close to 6900k performance (no idea if it will). So basically you have no faith in this CPU at all.

You want this CPU to perform well...why? Because from Sandy Bridge to Kaby Lake...Intel has become lazy and we haven't got that much of an IPC increase. And Intel still charges $350 for a 4c/8t part. https://hardforum.com/threads/kaby-lake-7700k-vs-sandy-bridge-2600k-ipc-review-h.1922342/

If AMD can release a $200 4c/8t CPU that is around the same performance as Kaby lake it is a win/win for everyone.

 

[Bandalo]

So basically you don't think games take advantage of more cores/thread (new games do), and you do not think it will come close to 6900k performance (no idea if it will). So basically you have no faith in this CPU at all.

You want this CPU to perform well...why? Because from Sandy Bridge to Kaby Lake...Intel has become lazy and we haven't got that much of an IPC increase. And Intel still charges $350 for a 4c/8t part. https://hardforum.com/threads/kaby-lake-7700k-vs-sandy-bridge-2600k-ipc-review-h.1922342/

If AMD can release a $200 4c/8t CPU that is around the same performance as Kaby lake it is a win/win for everyone.

I don't think Intel has gotten lazy, I think there's just not much gas left in the x86 tank at this point. To significantly improve IPC, they need to significantly change the architecture. And honestly for 90% of their sales there's no REASON for faster CPUs. Home and business users aren't clamoring for more performance, so why spend R&D funds to deliver it? If they dropped a chip with twice the performance of Kaby Lake tomorrow, most home and business users would never notice the difference.

Intel charges what the market will bear. Don't think AMD won't do the same, except maybe a small undercut in price.

 

[Deleted member 126051]

Going to wait and see. The last truly worthwhile offering (for me) was the Athlon64 Dual-Core.

AMD's going to need to put out something with the proper balance of price and performance to woo me away from Intel, even if Intel's gotten lazy over the last couple generations.

 

[Gigantopithecus]

Definitely buying, and the rest of the computer (except the HSF) was ordered yesterday. Whether I keep it or sell it, remains to be seen based on its performance. From what we've seen so far, officially and unofficially, it will likely perform slightly worse than my OC'd 4790K in single-threaded tasks and substantially better in multi-threaded tasks. If this is true, I will sell my 4790K system, as my workflow is more dependent upon the speed of my multi-threaded tasks.

 

[Dayaks]

If AMD can match Intel's performance at substantially lower price I might. It has to be inviting enough to outweigh the risks.

I have the itch to build a SFF.

 

[Riptide_NVN]

I was leaning towards Zen or Broadwell-E but now I'm not so sure.

I still can't decide if I should go with higher clocked 4c/8t CPUs or go for a couple more cores at the expense of clock speed. For my use (games, htpc, email/web) a 4c/8t CPU would be the best choice right now but what about 2-3 years down the road. Maybe the extra cores would have made more sense.

 

[walwalka]

I'm still waiting until official launch reviews, but I'm excited to see what they have done. So far, everything looks solid.

It will, very much be dependent on price.

 

[drescherjm]

I still can't decide if I should go with higher clocked 4c/8t CPUs or go for a couple more cores at the expense of clock speed. For my use (games, htpc, email/web) a 4c/8t CPU would be the best choice right now but what about 2-3 years down the road. Maybe the extra cores would have made more sense.

Lets hope for a 6C /12T Zen.

 

[Tzeh Pesh]

Wait and see for me.

Currently running a i5 760 with some regrets at not having held out for Sandy Bridge way back when and the upgrade bug has been biting pretty hard. Figure will likely end up with either Kaby Lake or Ryzen but having waited this long may as well wait a little longer to avoid a repeat. While gaming has been the most taxing thing I do these days, like the idea of potentially getting something with 6C/12T without having to necessarily go into enthusiast level parts/pricing, but get the feeling that's probably a little optimistic at least for this round of competition (2018 the year we start moving up from 4C/4T and 4C/8T at the consumer level?).

Seeing the Blender performance increases between series in the Kaby Lake 7700K vs Sandy Bridge 2600K IPC Review compared to some of the other benchmarks used, particularly single thread, has me a little wary of drawing conclusions from what's been presented so far, but am eager to see independent reviews and how the new platform fairs when people start using it day-to-day... while seeing the performance is probably my main reason for waiting, also a little wary of any bugs or kinks that might need to be ironed out when dealing with a new platform that I would much rather avoid having to deal with.

Definitely not swayed by the idea of either "punishing" Intel for their small performance increases of late or "congratulating" AMD for having finally managed to play catch up by going out and buying Ryzen just for the sake of it if it comes close... if the performance increases and price points aren't there, may end up skipping both Kaby Lake and Ryzen.

</admittedlyignorantopinions>

 

[Shintai]

If AMD can release a $200 4c/8t CPU that is around the same performance as Kaby lake it is a win/win for everyone.

Why would AMD do this? AMD isn't some kind of charity pro bono style organization.

 

[Shintai]

I was leaning towards Zen or Broadwell-E but now I'm not so sure.

I still can't decide if I should go with higher clocked 4c/8t CPUs or go for a couple more cores at the expense of clock speed. For my use (games, htpc, email/web) a 4c/8t CPU would be the best choice right now but what about 2-3 years down the road. Maybe the extra cores would have made more sense.

I dont think anything changes core wise the next 2-3 years and longer. Consoles are more or less limited to 6 cores as well, with the huge cluster penalty of ~190 cycles in between so using more than 4 got a huge limitation of its own. And even in DX12 games i3 and now KBL Pentium still beat much higher core parts. High IPC+high frequency will still rule gaming.

 

[JustReason]

I dont think anything changes core wise the next 2-3 years and longer. Consoles are more or less limited to 6 cores as well, with the huge cluster penalty of ~190 cycles in between so using more than 4 got a huge limitation of its own. And even in DX12 games i3 and now KBL Pentium still beat much higher core parts. High IPC+high frequency will still rule gaming.

where is this proof of i3 and Pentium. Most graphs of FPS show it far off from high core parts. Hell some games wont even play on dual core parts.

 

[Shintai]

where is this proof of i3 and Pentium. Most graphs of FPS show it far off from high core parts. Hell some games wont even play on dual core parts.

Not running on 2 cores / 2 threads is purely artificial due to consoles locking 0 and 1. And most of the few games, if not all of the games in question have been patched to fix this since release since it also affects performance on higher core/thread parts.

https://www.computerbase.de/2017-01/intel-pentium-g4560-test-kaby-lake/
http://gamegpu.com/action-/-fps-/-tps/deus-ex-mankind-divided-test-gpu
http://gamegpu.com/action-/-fps-/-tps/mafia-iii-test-gpu
http://gamegpu.com/action-/-fps-/-tps/titanfall-2-test-gpu
http://gamegpu.com/action-/-fps-/-tps/call-of-duty-infinite-warfare-test-gpu
http://gamegpu.com/action-/-fps-/-tps/dishonored-2-test-gpu
http://gamegpu.com/action-/-fps-/-tps/watch-dogs-2-test-gpu
http://gamegpu.com/mmorpg-/-онлайн-игры/ark-survival-evolved-test-gpu
http://gamegpu.com/rts-/-стратегии/sid-meier-s-civilization-vi-test-gpu
http://gamegpu.com/rts-/-стратегии/total-war-warhammer-test-gpu
http://gamegpu.com/rts-/-strategii/xcom-2-test-gpu.html
http://gamegpu.com/rts-/-стратегии/cossacks-3-test-gpu

Sum of all 2016 games tested.
http://gamegpu.com/test-video-cards/opredelenie-naibolee-bystrykh-i-optimalnykh-reshenij-2016-goda

 

[OrangeKhrush]

They d

Why would AMD do this? AMD isn't some kind of charity pro bono style organization.

they did sell a RX480 with 980 performance at $200 so a $200 4C/8T is possible.

Pricing will become apparent soon enough

 

[Shintai]

They d
they did sell a RX480 with 980 performance at $200 so a $200 4C/8T is possible.

Pricing will become apparent soon enough

Yes and NVidia sells 980 performance as Pascal for what? Fiji pricing, FX9xxx pricing, Kaveri pricing. There are plenty of examples. If a 4C/8T is 200$ its because there is ~200$ worth of performance in its metrics.

 

[Magic Hate Ball]

Yes and NVidia sells 980 performance as Pascal for what? Fiji pricing, FX9xxx pricing, Kaveri pricing. There are plenty of examples. If a 4C/8T is 200$ its because there is ~200$ worth of performance in its metrics.

There is a simple economic reason for why they could sell a 4C/8T chip for $200.

All of the i5 series chips have an iGPU attached that wastes die space if you're not using it. That costs money to manufacture, and it's around A THIRD (edit, not a half) of the die space of an i5/i7 chip. Therefore, a pure CPU part that's just 4C/8T with the Ryzen design (non-APU) will be much smaller and therefore cheaper to manufacture.

This will allow them to be very price competitive in the i3/i5 gaming on the "cheap" demographic.

EDIT: AMD usually goes with denser designs, which means it could be even smaller than a pure CPU part from Intel to boot.

And here's a helpful diagram

 

[Shintai]

There is a simple economic reason for why they could sell a 4C/8T chip for $200.

All of the i5 series chips have an iGPU attached that wastes die space if you're not using it. That costs money to manufacture, and it's around HALF the die space of an i5/i7 chip. Therefore, a pure CPU part that's just 4C/8T with the Ryzen design (non-APU) will be much smaller and therefore cheaper to manufacture.

This will allow them to be very price competitive in the i3/i5 gaming on the "cheap" demographic.

EDIT: AMD usually goes with denser designs, which means it could be even smaller than a pure CPU part from Intel to boot.

There is currently only 1 Zen die, its 8C/16T and you have to wait for the APU for a 4C/8T it seems.

Not having an IGP is part of the problem. Since the target market for the chip is so much smaller, hence increasing cost per unit. Secondly AMD is not an IDM either so GloFo takes their share of the cake.

 

[Magic Hate Ball]

There is currently only 1 Zen die, its 8C/16T and you have to wait for the APU for a 4C/8T it seems.

Not having an IGP is part of the problem. Since the target market for the chip is so much smaller. hence increasing cost. Secondly AMD is not an IDM either so GloFo takes their share of the cake.

AMD has stated in interviews that there will be a full product line up at launch. So, yes, there is no APU yet for cheap desktops, but their enthusiast/workstation space (which I think we fall into) will be using graphics cards paired with CPUs.

They will likely have harvested dies to make 4C (or 6C) parts at first.

 

[drescherjm]

All of the i5 series chips have an iGPU attached that wastes die space if you're not using it. That costs money to manufacture, and it's around A THIRD (edit, not a half) of the die space of an i5/i7 chip.

For Intel the cost savings of not including the iGPU may not be worthwhile. Remember that 90+% of users will use the iGPU.

 

[Shintai]

AMD has stated in interviews that there will be a full product line up at launch. So, yes, there is no APU yet for cheap desktops, but their enthusiast/workstation space (which I think we fall into) will be using graphics cards paired with CPUs.

They will likely have harvested dies to make 4C (or 6C) parts at first.

Its very unlikely you ever see a 4C/8T native CPU from AMD. It doesn't make sense from a business perspective. CPU product only in terms of enthusiast/workstation is a very small product. It needs to sell in the server space to be worth it. And that´s what Zen is, a server product first.

The 8C/16T die is estimated to be around 180-200mm2. 4+2 SKL/KBL with IGP is ~120mm2.

 

[Gideon]

I dont think anything changes core wise the next 2-3 years and longer. Consoles are more or less limited to 6 cores as well, with the huge cluster penalty of ~190 cycles in between so using more than 4 got a huge limitation of its own. And even in DX12 games i3 and now KBL Pentium still beat much higher core parts. High IPC+high frequency will still rule gaming.

Using more cores is the future in gaming and it's pretty obvious that quad core will be the least you want and people with more cores will get a boost. Each core can be dedicated to many things allowing the developers to have more power without making the game more efficient in the use of the processor. The biggest limitation has been the fact that Intel only sells quad cores and dual cores for the mainstream and hex cores and up to the high end and most developers wont code for things until it's more mainstream. If was not for AMD I doubt you would see newer games using more than 4 cores. If Zen sells well I expect games to use more cores and quads to become the minimum needed to run it. Dual core is a waste and I think Intel knows this as they are now putting out a dual core Hyper Threaded chip. The future is more cores as we have hit the limit on mhz and IPC for now. AMD was jut too early the first time with more cores and sacrificed too much IPC for it.

 

[Shintai]

Using more cores is the future in gaming and it's pretty obvious that quad core will be the least you want and people with more cores will get a boost. Each core can be dedicated to many things allowing the developers to have more power without making the game more efficient in the use of the processor. The biggest limitation has been the fact that Intel only sells quad cores and dual cores for the mainstream and hex cores and up to the high end and most developers wont code for things until it's more mainstream. If was not for AMD I doubt you would see newer games using more than 4 cores. If Zen sells well I expect games to use more cores and quads to become the minimum needed to run it. Dual core is a waste and I think Intel knows this as they are now putting out a dual core Hyper Threaded chip. The future is more cores as we have hit the limit on mhz and IPC for now. AMD was jut too early the first time with more cores and sacrificed too much IPC for it.

The exact same was said at the Bulldozer launch. Yet you can see the 2016 games benchmark.

If you look at Steam, then there isn't even half a million with 8 or more cores.

 

[Gideon]

The exact same was said at the Bulldozer launch.

And as I said they were too early and hurt themselves by lowering IPC too much for better multicore performance. But AMD has to capture a decent share of the market to make more developers take advantage of it. Right now AMD holds too little market share for a developer to say lets code for more cores. I am going to assume most people have a dual core and a few have quad cores right now. If ZEN sell well then developers will start assuming more people have more cores to put to use.

 

[Magic Hate Ball]

The exact same was said at the Bulldozer launch. Yet you can see the 2016 games benchmark.

If you look at Steam, then there isn't even half a million with 8 or more cores.

Mostly because AMD's Bulldozer and its descendants were performance failures relative to Intel. Not to mention getting stuck on older process nodes hurt their efficiency vs Intel as time went on.

Look, I'm not saying this is a sure thing at all, but I remember when AMD upset Intel with the Athlon, and then again with AMD64. It's possible. And if this is primarily a server chip, I can tell you a hell of a lot of them are high margin parts with no iGPU, in a market they can only go upwards in marketshare.

 

[Shintai]

And as I said they were too early and hurt themselves by lowering IPC too much for better multicore performance. But AMD has to capture a decent share of the market to make more developers take advantage of it. Right now AMD holds too little market share for a developer to say lets code for more cores. I am going to assume most people have a dual core and a few have quad cores right now. If ZEN sell well then developers will start assuming more people have more cores to put to use.

Like on the 8 core consoles?

There is a reason why more cores isn't used. And its not due to lacking CPUs with more cores.

 

[Magic Hate Ball]

Like on the 8 core consoles?

There is a reason why more cores isn't used. And its not due to lacking CPUs with more cores.

There are some games that are using more cores reliably. Overwatch for example can be shown in this test to skew heavily on 4C/8T i7's versus 4C/4T i5's, even if the i5 is higher clocked http://www.techspot.com/review/1180-overwatch-benchmarks/page5.html

Pay attention to the gap between the i7-4770K and the i5-4690K. They are both Haswell, 22nm chips but perform VERY differently in this test despite identical clock and boost numbers (stock).

i7: 175 minimum - 232 average FPS
i5: 123 minimum - 180 average FPS

 

[Gideon]

Like on the 8 core consoles?

There is a reason why more cores isn't used. And its not due to lacking CPUs with more cores.

Consoles are a whole different game and they use all those cores on the high end games. I would not use consoles as a tool to compare against desktop computers and how games utilize them.

 

[Shintai]

There are some games that are using more cores reliably. Overwatch for example can be shown in this test to skew heavily on 4C/8T i7's versus 4C/4T i5's, even if the i5 is higher clocked http://www.techspot.com/review/1180-overwatch-benchmarks/page5.html

Pay attention to the gap between the i7-4770K and the i5-4690K. They are both Haswell, 22nm chips but perform VERY differently in this test despite identical clock and boost numbers (stock).

i7: 175 minimum - 232 average FPS
i5: 123 minimum - 180 average FPS

5960X 213 FPS.
FX4320 145 FPS.
FX6350 143 FPS.
FX8370 146 FPS.
7870K 72 FPS ??

Note no scaling for FX CPUs. And Kaveri just performing awful.

I would like to see more benchmarks and patched game before calling that one.

 

[Pusher of Buttons]

I'll buy ryZen in a heart beat...but the issue AMD has had lately (in addition to an obviously dated chip architecture) is the lack of solid motherboards. They essentially don't exist in the ITX space for AM3 boards. So...if they can get the mobo builders to produce solid products, and Ryzen winds up being as good as advertised, of course I'l be a customer.

 

[Magic Hate Ball]

5960X 213 FPS.
FX4320 145 FPS.
FX6350 143 FPS.
FX8370 146 FPS.
7870K 72 FPS ??

Note no scaling for FX CPUs. And Kaveri just performing awful.

I would like to see more benchmarks and patched game before calling that one.

Look, we know the FX CPU's aren't good for gaming vs i5/i7's. They also split FP units (not true full 8C), and have all sorts of issues, so you're seriously throwing a red herring out there by including them.

I'm looking at two nearly identical chips, except one is 4C/8T and the other is 4C/8T and you're not convinced.

I led the horse to water and it didn't drink.

The 5960X is significantly lower clocked, and still beats the i5.... not sure how you're going to call that a loss for me?

 

[OrangeKhrush]

Y

Its very unlikely you ever see a 4C/8T native CPU from AMD. It doesn't make sense from a business perspective. CPU product only in terms of enthusiast/workstation is a very small product. It needs to sell in the server space to be worth it. And that´s what Zen is, a server product first.

The 8C/16T die is estimated to be around 180-200mm2. 4+2 SKL/KBL with IGP is ~120mm2.

AMD have a 4/8 SoC ravenridge on slides and Canard confirmed a bottom end summit Ridge 4/4 Ryzen CPU and more surprising at clocks and TPD you get in workstations.

I think the SKU lineup will be somewhere like

SR7 black Edition 8/16
sR7
sr5 6/12
sr3 black Edition 4/8
sR3 4/4

Performance will be priced around what Haswell performance would cost.

 

[Shintai]

AMD have a 4/8 SoC ravenridge on slides.

Raven Ridge is a H2 2017 APU product.

 

[Shintai]

Look, we know the FX CPU's aren't good for gaming vs i5/i7's. They also split FP units (not true full 8C), and have all sorts of issues, so you're seriously throwing a red herring out there by including them.

I'm looking at two nearly identical chips, except one is 4C/8T and the other is 4C/8T and you're not convinced.

I led the horse to water and it didn't drink.

The 5960X is significantly lower clocked, and still beats the i5.... not sure how you're going to call that a loss for me?

So you decide to completely dismiss the FX results. 2M4T, 3M6T, 4M8T all performed the same at same clocks. The 5960X vs a 3Ghz SKL raises its own questions. There are no benefit with real cores vs HT.
The Kaveri chip in the test performed half of the 2M4T FX. Lack of L3 cache? The HT parts also got 2MB more cache than the non HT parts if that got any effect.

There is a reason why I asked you for numbers after patches etc.

 

[OrangeKhrush]

Actually a 6600K beats a 5960

Look, we know the FX CPU's aren't good for gaming vs i5/i7's. They also split FP units (not true full 8C), and have all sorts of issues, so you're seriously throwing a red herring out there by including them.

I'm looking at two nearly identical chips, except one is 4C/8T and the other is 4C/8T and you're not convinced.

I led the horse to water and it didn't drink.

The 5960X is significantly lower clocked, and still beats the i5.... not sure how you're going to call that a loss for me?

Actually a 6600K can beat a 5960

 

[OrangeKhrush]

The 1100T was far superior to Zambezi, i had an FX for a day before going back to an old faithful. I moved my HTPC, workbench and SFF to intel in 2013.

I found a compiled list of the Blender scores from a whole plethora of CPU's,. The outcome was less SMT related and more architectural on careful break down of the scores. Interestingly a 1100T @4ghz handsomely beat up a 4.8ghz 8370.