Intel Broadwell Thread

[RussianSensation]

The big downside to a May/June launch (which seems reasonable enough) is that this probably means SKL-E is a May/June 2017 affair. BDW-E being more compelling than thought should help mitigate that though.

Wow, whoever bought a 5820K in September 2014 is smiling right now. That means it might be almost 3 years before there is a somewhat justifable upgrade path for them, unless they decide to jump on the 8-10 core BW-E. Seems like SLK-E for mid-2017 at the earliest is starting to become a real possibility with BW-E delays. The sad part is that by the time SKL-E launches, Skylake architecture will be approaching 2 years old. Another kick in the nuts for the HEDT platform. At this pace it Icelake-E could be almost half a decade away.

Even then, it might be arguable if SKL-E is even a good upgrade path over 5820K without going with more cores given that SKL only offers about a 10% increase in IPC over HW (when averaging games+applications). Certainly whoever bought the 5820K in 2014 made a slam dunk decision since it might be 4-5 years out before there is any great upgrade path.

 

[jpiniero]

Actually it appears that Intel is splitting HEDT/Xeon E5 1P from the 2P+ models starting with Purley, not to mention diverging the core itself from the mainstream line. What that ends up meaning is unknown, but it's possible that Cannonlake-E could come earlier in the cycle but is more like a mainstream CPU without the PCH but more cores, quad channel memory, etc..

 

[tenks]

Wow! The members of the IST brigade decides that attack is the best defense. Better to attack the messenger for telling the truth instead of admitting you were wrong.

Hahaahha .... you are hilarious...

you need to get a life

 

[Fjodor2001]

you need to get a life

Did your feelings get hurt? It would be much easier for you to just admit you were wrong.

 

[Fjodor2001]

Also, I'm pretty sure those users will be running Broadwell-E long before you ever touch a Zen CPU.

Is there anyone who has claimed otherwise?

But that was not the issue here. This was about the latest Broadwell-E delays. Is it so hard for you to admit you were wrong about that? Just like you were wrong about Broadwell-E being 12 cores?

 

[Aristotelian]

Wow, whoever bought a 5820K in September 2014 is smiling right now. That means it might be almost 3 years before there is a somewhat justifable upgrade path for them, unless they decide to jump on the 8-10 core BW-E. Seems like SLK-E for mid-2017 at the earliest is starting to become a real possibility with BW-E delays. The sad part is that by the time SKL-E launches, Skylake architecture will be approaching 2 years old. Another kick in the nuts for the HEDT platform. At this pace it Icelake-E could be almost half a decade away.

Even then, it might be arguable if SKL-E is even a good upgrade path over 5820K without going with more cores given that SKL only offers about a 10% increase in IPC over HW (when averaging games+applications). Certainly whoever bought the 5820K in 2014 made a slam dunk decision since it might be 4-5 years out before there is any great upgrade path.

Probably! But the truth is on an overclocked 2600k (mildly - 4.4ghz) I don't really feel the CPU as sluggish - more the SSDs compared to modern systems I've used, etc. The only reason I'm growing antsy with any Broadwell-E delays is because I'm waiting to build an entirely new setup around it + next gen video cards.

And, to be blunt (this isn't directed towards you, RS) I'm surprised in the CPU forum to see threadcrapping mocking one side or the other for delays. Intel's CPUs have been doing well for so long - I'm happy to wait for a delay if it ensures the great experience I've come to expect from their products.

From an overclocked 2600k to an overclocked Broadwell-E 10 core (plus an entirely new system) - let's just say I'm excited

 

[mattgo]

I'm waiting,

 

[Sweepr]

Is there anyone who has claimed otherwise?

But that was not the issue here. This was about the latest Broadwell-E delays. Is it so hard for you to admit you were wrong about that?

For the record, I have been saying late Q1 or Q2 for Broadwell-E for some time now. I found it strange that some users were still expecting an early Q1 launch when Bench-life indicated Q2 (before the latest leaks). Some confusion might come from the fact that server models are actually launching earlier next year.

Just like you were wrong about Broadwell-E being 12 cores?

Let me expose your lies once again. This is what I said about Skylake-E, not Broadwell-E:

There are scarce details about core count, clockspeeds, etc. For all we know it might offer 10-12C at the high-end ($1000), especially if AMD delivers. Servers are getting 55% more cores from Haswell to Skylake.

So I was right (and you were wrong again), Broadwell-E already delivers 10C/20C, Skylake-E is either another ten-core chip or something better.
Now that we've cleared that can you please provide the benchmark results (and price/launch info) you used to come up with this?

You're just bitter because in 2016 you'll be sitting on an expensive and slow 4 core Intel CPU, while others will be using a cheaper and faster 8 core AMD CPU.

We're still waiting.

 

[Sweepr]

From an overclocked 2600k to an overclocked Broadwell-E 10 core (plus an entirely new system) - let's just say I'm excited

Talk about an upgrade.
I'm still rocking 6C/12T Gulftown in one of my machines, almost 6 years and counting (also a 7 year old 4C/8T Bloomfield!). I feel like Haswell-E/Broadwell-E are going to be the new kings of longevity from the Intel camp.

 

[IndyColtsFan]

Talk about an upgrade.
I'm still rocking 6C/12T Gulftown in one of my machines, almost 6 years and counting (also a 7 year old 4C/8T Bloomfield!). I feel like Haswell-E/Broadwell-E are going to be the new kings of longevity from the Intel camp.

I agree. If I upgrade my rig to Broadwell E (or even do a quick and cheap build with Haswell E and then upgrade to a larger core count later via Broadwell E), I wouldn't be terribly surprised if it is a viable system for 10 years assuming that I do regular GPU updates. I even think my 2600K could work well for me for another 3 years, but the upgrade itch needs to be scratched soon.

 

[Fjodor2001]

We're still waiting.

It's based on guesstimates from other forum members, and you can continue waiting for ~1 year until Zen is expected to be released. But you've asked the same question like 1000 times already, so you already know this and are just trying to create drama and divert from the original issue as usual.

 

[tenks]

Did your feelings get hurt? It would be much easier for you to just admit you were wrong.

What about every SINGLE post you make about Zen? Where's that awesome lower cost, higher performing AMD 8core you said you were going to be using in 2016?

Exactly. Hypocrite.

And the only drama this forum has is you creating it. Look around man, we were talking about the topic instead you came in and derailed it, as usual. You really are a troll

 

[Fjodor2001]

What about every SINGLE post you make about Zen? Where's that awesome lower cost, higher performing AMD 8core you said you were going to be using in 2016?

Exactly. Hypocrite.

And the only drama this forum has is you creating it. Look around man, we were talking about the topic instead you came in and derailed it, as usual. You really are a troll

The original issue was about Broadwell-E getting delayed. Then the IST brigade went into denial and misinformation mode, trying to derail it into a Zen story (!). Thereby trying to divert from the original issue and avoiding having to admit that their earlier denial of Broadwell-E being delayed proved to be incorrect.

You really are ridiculous. I'd be surprised if you're even believing what you write yourself.

 

[Arachnotronic]

The original issue was about Broadwell-E getting delayed. Then the IST brigade went into denial and misinformation mode, trying to derail it into a Zen story (!). Thereby trying to divert from the original issue and avoiding having to admit that their earlier denial of Broadwell-E being delayed proved to be incorrect.

You really are ridiculous. I'd be surprised if you're even believing what you write yourself.

LOL, did you even read Sweepr's posts? He was telling us that leaks said Q2 even while folks like me (based on BDW-EP roadmap leaks) were saying Q1.

By the way, totally admit that I was wrong on the Broadwell-E launch timeframe. It's a bummer, but at least it is a product I and others actually want thanks to the 8 core coming down in price compared to the HSW-E 8 core and the 10 core taking the top EE spot.

Sweepr is easily one of the most intellectually honest posters on this board and I think that you destroy your own credibility by labeling him a mindless Intel shill by categorizing him as part of your (fictional) "Intel Sales Team."

 

[coercitiv]

Sweepr is easily one of the most intellectually honest posters on this board

+1 for Sweepr from me too, and I'm one of the posters who likes to nitpick

Can we get back to some better topic on the Broadwell thread though? Considering Skylake had less trouble with overclocking than "desktop" Broadwell, would you say BW-E will be a decent overclocker?

 

[Fjodor2001]

LOL, did you even read Sweepr's posts? He was telling us that leaks said Q2 even while folks like me (based on BDW-EP roadmap leaks) were saying Q1.

Yes, now that it is obvious and undeniable. But not before when this was discussed earlier and there were leaks about Broadwell-E being delayed.

Sweepr is easily one of the most intellectually honest posters on this board and I think that you destroy your own credibility by labeling him a mindless Intel shill by categorizing him as part of your (fictional) "Intel Sales Team."

If you've failed to see his unbalanced bias towards Intel then I think you need to re-read his posts.

 

[Fjodor2001]

Considering Skylake had less trouble with overclocking than "desktop" Broadwell, would you say BW-E will be a decent overclocker?

Perhaps, if the reason Skylake had less trouble overclocking was because the 14 nm process is more mature now. But if it was due to some uArch or design changes from Broadwell->Skylake, then Broadwell-E should have similar OC problems as desktop Broadwell.

But is it an accepted fact that Skylake overclocks much better than Broadwell to begin with? The problem is also that there are no equivalent desktop SKUs to compare. Broadwell only has the "specialized" 5675C/5775C CPU variants with Iris Pro iGPU.

 

[ShintaiDK]

The only Broadwells overclocked was GT4e parts with EDRAM. That messes up the result big time in a direct compare.

 

[dark zero]

So.... Broadwell E got delayed again?

 

[witeken]

So.... Broadwell E got delayed again?

Possibly for protecting gross margins: http://intelstudios.edgesuite.net/im/2015/pdf/2015_InvestorMeeting_Stacy_Smith_WEB4.pdf (slide 28-30).

 

[DrMrLordX]

The only Broadwells overclocked was GT4e parts with EDRAM. That messes up the result big time in a direct compare.

The i7-5775c has gone as high as 4.8 GHz on a good AIO cooler. Fugger required a fair amount of voltage to get there (~1.5-1.55v if I recall, which isn't so extreme for Intel 14nm). I don't know that Broadwell-E will have lesser clockspeed potential than Haswell-E.

 

[tenks]

I've always said late March 2016. At that time, that was the date it was delayed to in leaked roadmaps. You made a blanket statement, so if it got delayed again you could just be like "HERP DERP" like you're doing now. Man, you totally showed me and everyone else who said March q2. Now it's May 31st..60 days later. Game changing stuff. Meanwhile, where's Zen? oh right..

It makes sense why you're doing this though..you're an AMD fanboy with a huge void to fill so you're just bagging on anything Intel. You trolling won't help, as that void is pretty dark and lonely.

So.... Broadwell E got delayed again?

Broadwell-E never got an official launch date by Intel when they announced the first delay, so technically, it's on time. But it's a couple months later then what everyone speculated.

When there is absolute 0 competition from anyone but themselves, Intel has no reason to rush. The EP's have a Q1 release, I'm sure if they wanted they could drop the HEDT sku's sooner.

 

[witeken]

When there is absolute 0 competition from anyone but themselves, Intel has no reason to rush. The EP's have a Q1 release, I'm sure if they wanted they could drop the HEDT sku's sooner.

The reason it got delayed has everything to do with 14nm yields. And spoiler alert:

1. There's plenty enough competition with Samsung and TSMC
2. The theoretically best thing one could do is pursue Moore’s Law at infinitesimal intervals between nodes, so going faster to the next generation is always better.
3. The thing that investors want Intel to do, is get the yields as high as possible, since that will give the highest margins... Obviously. So no company would ever intentionally cripple their own yields. So when you look at the latest yield curve, it wasn't nearly as steep as Intel expected. Starting Q4'14, the pace of improvement slowed down considerably, even though by then they were about on par with 22nm, but they fell again behind.

 

[tenks]

The reason it got delayed has everything to do with 14nm yields. And spoiler alert:

1. There's plenty enough competition with Samsung and TSMC
2. The theoretically best thing one could do is pursue Moore’s Law at infinitesimal intervals between nodes, so going faster to the next generation is always better.
3. The thing that investors want Intel to do, is get the yields as high as possible, since that will give the highest margins... Obviously. So no company would ever intentionally cripple their own yields. So when you look at the latest yield curve, it wasn't nearly as steep as Intel expected. Starting Q4'14, the pace of improvement slowed down considerably, even though by then they were about on par with 22nm, but they fell again behind.

1) producing high-end 10core x86 chips on 14nm? Yea, no.

 

[Sweepr]

Just in case some people forget, here's some Core i5-5675C gaming results at 1080p.

http://fcenter.ru/online/hardarticles/processors/38489-Obzor_processora_AMD_A10_7870K_Godavari

http://itc.ua/articles/14-nanometrov-dlya-nastolnyih-pk-obzor-protsessora-intel-core-i7-5775c

Also the performance it provides in CPU-limited games when coupled with a dGPU. Too bad the limited availability in some countries pushed the prices up.

Looking forward to the Skull Canyon NUC running Skylake GT4e in Q1-2016.