Is that a Caterpillar or a Daewoo??
Definitely not Daewoo. Close in appearance to Caterpillar, though there are numerous companies using orange-yellow paint schemes for their bulldozers.
I hope you’re not planning on getting the Religious Right all a-flame by following up in January with a strip showing the Trinity being blown up….
No need for religious themes, there’s a much better source for such future comic strip:
Nice attempt, but would be even better if teh bulldozer would be standing in front of something like this:
which would still look intact in the last image of the strip
I got this one! I thought it was humorous enough.
It’s not an explosion, but a transformation. (like Gandalf the White kind of thing, where you can’t see it directly)
The biggest problem with Bulldozer is that AMD marketing drones hype it to be a Sandy Bridge killer, when it clearly wasn’t capable of achieving that. Likewise, some computer enthusiasts were expecting it to put a dent in Sandy Bridge’s lead. The underlying design made it clear that Bulldozer meant for heavily multi-threaded workloads which only exist in the server arena.
AMD has to do a product refresh with their desktop line-up, whatever the product is up to the task or not. Intel did the same thing back the days when Netburst was getting pounded by the K8. It seems some people do not remember the overpriced, and lackluster Pentium 4 Extreme Edition chips.
At best, Bulldozer is just AMD’s two-year late answer to Bloomfield and its siblings. It has comparable overall performance and power consumption. Sandy Bridges are just too power efficient. They also have a slight performance boost over their Bloomfield predecessors.
I suspect Bulldozer’s lackluster FP performance is attribute to the fact that AMD’s architectures are expecting the integrated GPU to take over FPU’s traditional role. Intel is going to follow suit as well.
their low IPC is also deliberate in an attempt to fit 8 int cores into a single chip. as the manufacturing process gets better and they are able to put more transistors into a chip, the IPC will improve.
the need for high frequencies is for the uncore to work fast enough to pump instructions to both cores
overall its a good strategy, only problem is that intel is churning out amazing chips that goes beyond the relm of good into that of magical.
Krog is on point!
what people don’t see is that bulldozer follows the raedon architecture where key components are built into small modules that can be easily added or removed.
the FP modules will eventually be replaced with a full fledged functioning GPU that will bulldoze whatever FP module intel has. the integer modules will continue evolve into effective smaller modules that run efficient and fast but the modules will have to be packed smaller first to accommodate 8 of such modules into a single chip; thus the lack of IPC currently.
AMD isn’t intel, they don’t have the capacity to design a chip based on today’s manufacturing process, they have to design a chip based on manufacturing processes 2 or 3 times ahead of today’s.
thats why phenom failed and phenom2 based on a more advanced manufacturing process managed to match the core2quads. BD is the 1st step in a long line and probably wasn’t made for 32nm process thus sacrifices will have to be made. it’s concept is based on an evolutionary chip that is far ahead of its time and far ahead of today’s manufacturing process.
so lets not get ahead of ourselves, you don’t punish a baby for crawling do you?
When the baby is 42 years old and lives at home still, yes.
i’m not talking about x86, the BD architecture is only weeks old, not 42 yrs old.
give it time to grow, let amd put BD on a 22nm process, i’m sure that’ll fix every single constraint and bug BD has
The design measures its age in years at this point.
hardly. they have had working silicone fore over 2 years. and it still sucks.
“give it time to grow, let amd put BD on a 22nm process, i’m sure that’ll fix every single constraint and bug BD has”
That chip took forever to finally release… AMD can’t say it didn’t have enough time. It’s like a famished baby without a head popping out after a 36-month pregnancy!
How long exactly would you think AMD needs?
they need a better manufacturing process. BD can’t be made on 32nm, it has to be on 28 or even 22nm for it to show its true potential
It’s not an efficient design. Plain and simple. Would you also say that the Pentium 4 should have been made on a smaller process to show it’s potential? Just because the last batch of Pentium D’s were on a 65nm process that made them run cooler does not negate the fact that at the same time Core 2’s were running rings around them at lower clock speeds AND TDP on that same node.
A flawed design is a flawed design. You can’t make junk and then try to say it’s ahead of it’s time. Customers are looking for solutions in the present.
i am sure your trolling, but i cant help myself
as a owner of a i750@4ghz and a x61090t@4ghz… amd *never* caught up with intel, in the reality or any of your dreams.
lastly, intel desgins cpu’s 4 generations out, and also outsources 2-3 projects per cycle to see what those crazy israli teams and cook up. your clearly a marketer and need to be sexually abused by a yak.. go back to amd’s payroll troll. Even the dumbest TR troll here is brighter than you…
hell snakeoil might be smarter than you
i’ll bet my last dollar that phenom2 955 is wayy faster than any intel core2duo
who says amd never caught up with intel? you must be snakeoil
At four threaded encoding, sure.
Single or dual thread I don’t think AMD has caught C2D – at least 45nm, 3.3ghz C2D – yet and with Faildozer the gap is getting bigger instead of smaller.
WRONG :: [url<]https://techreport.com/articles.x/20873[/url<] not even in 20% range snakeoil
[quote<](#75 forumics), WRONG :: [url<]https://techreport.com/articles.x/20873[/url<] not even in 20% range snakeoil[/quote<]
Eh, what you are linking to doesn't show an X4 955, and the only Core2 Duo is the E6400.
I think forumics punked you.
doesn’t matter which phenom2 we are using against which core2duo, my point was that AMD caught up with intel before and will surely do it again, its just a matter of time
yeah and it took YOU (AMD) 6 god damn years to do it. want a gold star ???
Dude, that poster is a troll (and very poor at it, b.t.w), just ignore it :-/
Mmm, replying to me but using maxxcool’s quote. Yeah.
You seem to be stuck on comparing the Phenom II with the “core2duo”. Since most Phenom II’s are X4’s, so you should be comparing them to Intel’s Core2 Quad. In this case, by the time AMD (arguably) caught up to Core2 Quad, Intel already had Nehalem out.
well shit, i didn’t know we were going back in a time machine… besides. people compare modern to modern. hes talking about taking a evoiii against a 1965 mustang…
Heh, I was just bustin’ on ya a little because forumics slipped “core2duo” in there when the pertinent CPU to use when comparing to most Phenom II’s would have been “core2quad” (to use forumics’ gibberish), which is what I’m guessing you read.
We need more puns.
I’ll try to start:
The third image clearly shows the Bulldozer is quite “energetic”.
Ah yes, the pun. Where unfunny people go to laugh.
Momma jokes then?
Yo bulldozer, yo ‘ momma iz so fat she gave birth to a efing huge construction site machinery and she named you after a smaller one!
Bulldozers require juice, are noisy and slow.
So, AMD, develop a cheetah
My slightly altered caption:
Fredo, you are such a Papermaster…oops sorry Comicmaster!
B3 stepping… AMD catching the flying parts in the far right image to build a TATA…
Bobcat: Too slow for mainstream
Llano: Too Rare to find
Bulldozer: To be scraped.
Tata: Too late… ha ha ha
AMD: To be unannounced… hi hi hi
I don’t wish any ill towards any company, be it intel or amd, or even rambus (amazing, i know) since the more, the merrier.
but i gotta admit, this was funny as hell.
What a lame comic strip. Stop doing comics if you don’t have witty imagination.
Clearly comedy is hard.
and sarcasm is easy. or is it hard? oh no!
Spoil sport. I likes a good visual.
Granted, it would have been funnier if the explosion were throwing the machine apart into eight pieces, but we can’t have everything.
What a lame comment. Stop doing comments if you don’t have a witty imagination.
[i<]Edited 1 time(s). Last edit by your mom on Oct 25 at 06:04 AM.[/i<]
Ouch. But I think an explosion is the wrong image. Some steam and smoke would be more like it.
That said, the whole dozer disappointment is a bit of a shame for the customer, it would be better if AMD had a more competitive new architecture.
ADVANCED MICRO DISASTERS…
AMALGAMATED MACRO DOOMS…
ALARMING MICRO DISEASES…
The savior is coming… Trinity!!!! An up-scale disaster …..
Trin……TOCK…PONG,PONG……BANKRUPT. SUE INTEL AGAIN FOR CASH!
Yebo, why don’t you do one on Trinity Apocalypse??
ho hu hi ha ha ha
[quote<]ADVANCED MICRO DISASTERS...[/quote<]
That's actually very funny, imo.
“Bankcrupt”? What the heck is that?
The bulldozer was so heavy it fell off the sandy bridge. 😛
This is f***ing hilarious. Its still giving me a good laugh. Awesome comic.
The comic reminds me of the days when I used to drive inferior quality American cars before I switched to Honda.
That’s why the real overclockers use liquid nitrogen.
and only one module
Give me a Juniper for each core with a lucid thingy and we’re talking.
Hmm. Bulldozer or Lambo? With Ivy Bridge it’s probably more like bulldozer or Lexus LFA.
This is very funny, gave me a good laugh atleast.
Ok, now this one is funny!
They don’t run [i<]that[/i<] hot. Big and hot, like a bulldozer, but one more thing, how many people versus how many companies own bulldozers? This isn't designed to be a desktop chip.
Then why did they release it as a desktop chip touting it’s “gaming performance”?
Marketing did that, not the engineers or sites like TechReport.
No, the engineers put it in a desktop socket. If it was meant to be server-only, they’d have put it on socket G34 and C32.
But there is going to be a market for bulldozer on the desktop. It’s marketing that spouts the BS.
I’m going to stick my neck out here and say that Duck in some sense is correct. Only in the sense that BD does have a place on desktop. Especially if you use Linux. Linux is the one place where the kernel and the programs that support it are largely updated more frequently. It’s here where BD doesn’t score so badly, especially when video encoding.
Like here: [url<]http://www.phoronix.com/scan.php?page=article&item=amd_fx8150_bulldozer&num=10[/url<]
Also in Photoshop CS5.1 as seen here: [url<]http://www.tomshardware.com/reviews/fx-8150-zambezi-bulldozer-990fx,3043-15.html[/url<]
When was the last time AMD was competitive in Photoshop? Well never.
I definitely agree that BD is meant for server workloads. However, as desktop software moves to highly multithreaded integer workloads and largely that's the trend you will see, BD will start to perform like it should.
The downside to all of this is how many people are going to wait for that to happen??? Almost no one and this is the problem.
Who is going to buy BD because of Windows 8? Not many
Who is going to buy BD because of Linux? Maybe me and few others here, but they aren't the norm.
AMD is taking a risk here and a huge one that somehow the software will catch up in enough time to make BD viable for the masses. That's the one thing that I think will be BD's downfall. "If you build it they will come," has only worked in very few cases with AMD, X86-64 being one of them, but there's not too many examples of it working for AMD in others.
Would it have killed Phoronix to order the results? Would it? Hmm?
They sorted them, just not in a way that makes sense.
A lot of fair points.
Coupling those points with the fact that AMD has been hiring a *LOT* of Linux engineering talent speaks volumes about which market they want to target with BullDozer derived server/workstation parts and APUs for laptops/desktops/HTPCs.
Given that they have their own compiler now and that Linux customers traditionally have source code access, AMD can effectively offer in-house engineering expertise for customers who have special Linux performance needs, instead of having to go through several 3rd parties (e.g. MIcroSoft) to get effective performance tweaks implemented in a proprietary stack.
A recent example of this approach would be the the Linux scheduler patches awaiting inclusion in the mainstream kernel.
Time will tell whether this strategy will pay off down the line…
[quote<]Time will tell whether this strategy will pay off down the line...[/quote<]
By the time "down the line" comes around, AMD - a wholly owned subsidiary of XYZ Corp. - will have a different CPU architecture.
[quote<]When was the last time AMD was competitive in Photoshop? Well never. [/quote<]
No doubt, they improved in this metric, but I often times wonder how far to trust THG these days...
Most other sites that tested with CS5 it was matching to slightly behind the 2500K.
I would still call it competitive though. It is certainly the best showing in Photoshop from AMD yet.
Blaming “marketing” for a CPU’s under-performance is like blaming your Corolla’s paint color for it not being fast.
No matter how you spin it, we expect a next-gen CPU to be more powerful than it’s predecessor. That did not happen with Bulldozer.
No AMD as a company did that, you can’t just wave your hand and make decisions you don’t like go away.
What else could they do with what they had?
It’s clear BD isn’t for desktops, but it’s crazy how AMD insists that it is.
Both AMD and Intel are traversing a similar path. Smaller die, integrated GPU on die, simple two-chip platforms are for mainstream and OEM based systems. High end desktop, where discrete GPUs will be the norm derives a GPU-less CPU, large die, higher TDPs which are repurposed from the server products.
AMD would need to split out a separate design team, and target high end desktops with a different architecture, otherwise this argument “that it was really meant for servers” will always hang over the discussions.
[quote<]High end desktop, where discrete GPUs will be the norm derives a GPU-less CPU[/quote<]
Where's Intel's GPUless CPU? I'd love to have it, as my 2600k has a bunch of idle silicon in its 995 million transistors.
Oh... Bulldozer has somewhere around 2 billion, yet still falls short?
That just makes me wonder what a 2 billion transistor Ivy Bridge will pull off.
Yes it is. It’s a desktop chip, fool.
^Mr. T accent
Do you understand how humor works?
No, what’s that?
[quote<]This isn't designed to be a desktop chip.[/quote<]
If it's marketed to desktop users, then it's a desktop chip.
You’re right. It was apparently designed to be a space heater that could also run multi-user pong.
That reminds me, I miss Blast Corps for the N64.
Be epic. Be brutal. Experience the world’s first 8-core desktop processor.
Yeah right. Who are they kidding? Just so sad to see AMD resort to shady marketing stunts because they’ve failed spectacularly with arguably the most anticipated CPU launch ever.
At least it is the first 8-core desktop processor.
there are many camps that would debate that.
It’s not using SMT. Just because there’s not an FPU per core doesn’t mean that each core is not a core. Shared cache also doesn’t make each core not a core.
You can certainly argue against the fact that it’s two cores per module however.
In terms of the front-end, as with SMT, you only have single fetch and decode units, it’s only the integer scheduler and pipeline which gets doubled up – so if you can saturate the fetch and decode side with a single thread that second integer pipe can’t be utilised.
Likewise if you’re dealing with predominantly floating point code it’s little, if any, better than SMT.
I think AMD’s marketing team has it backwards.
When you have a good product, you advertise it as such, bring back the FX brand, etc.
If you have a crap product, don’t get too noisy, or hype it up. Just launch it, then people won’t be as disappointed.
I hope the Bulldozer based server chips are working out for Cray.
I can feel the heat from all the way over here!
BTW i noticed the duplication of the cogs and bolts, like the arhitecture, ha!
Sorry Fred; I am just sorry.
All Rights Reserved. Copyright Tech Report.