Still no announcements about Titan, decent rumors didn't start happening until this week, and release is supposed to be next week.
- Status
CurrentlyPissed
Senior member
- Feb 14, 2013
- 660
- 10
- 81
Meh, true I suppose. I just don't see it happening with the multiple statements, nothing on the roadmap, and the above link I posted as well.
Tweak155
Lifer
- Sep 23, 2003
- 11,449
- 264
- 126
What is it about this thread that causes a next page button to appear (I.E right now I'm on page 43, but 44 is in the page list) yet when I click page 44, it just takes me to the top of 43. This is the only thread I've seen this happen on this forum.
Some other guy with the same number of posts as you only 8 weeks ago said GK110 would never be a GeForce product. 20nm gpu's probably aren't coming until q2 of 2014, so it would make way more sense to refresh products now in between this current gen and 20nm than wait until the end of 2013, 5-6 months before 20nm gpu's come out. I can keep listing more reasons and indications that refreshes are coming sooner than later, but common sense is hard for some to fathom.
Last edited:
CurrentlyPissed
Senior member
- Feb 14, 2013
- 660
- 10
- 81
Ghost page, seen it happen on alot of VB boards. Not sure how it gets resolved though.
CurrentlyPissed
Senior member
- Feb 14, 2013
- 660
- 10
- 81
I don't know if you are saying him, and me are the same. Or what not, but I've actually frequented this forum for some time. I primarily post on Rage3D(same name, around 9k post over there) however. Just made an account over here to spread the love. So, no. I'm not some random fud spreader. Just keeping the conversation alive as am excited about the Titan release. But if it's what current rumors are, I may just go 7970 CFX. Or the new Club3D(powercolor) 7990 that's all the rave.
Not at all insinuating you are the same person, didn't mean that at all. My apologies.
On subject, Kepler is an established architecture. Making tweaks to it won't take as much resources as it would take to augment or overhaul an existing architecture. Fermi was refreshed in 6-8 months (depending on which chip in Fermi came out when). Kepler chips will get similar, albeit less significant refreshes about a year (give or take six weeks) after their original releases. That will put the refreshes squarely in between Kepler gen 1 and Maxwell. It will also allow nvidia to keep their prices similar to what they are at now without needing to heavily promote sales with game bundles.
Waiting until q4 to simply update existing chips 6 to 9 months before 20nm gpu's come out doesn't make any sense logistically and financially.
DiogoDX
Senior member
- Oct 11, 2012
- 757
- 336
- 136
What scares me in this topic is that some people are happy with the creation of the series EXTREME by Nvidia. I can see a future where Nvidia will release their major GPUs for $1000 and fill the rest of the market with cut down versions (aka GK104).
This. And the problem is that what sells for $1000 in USA sells for $1500ish in India and usually I prefer GPUs within $600-800 in India or US counterparts should costs sub $500-600 tops. That too is a bit high to change yearly but more than twice of that is really excessive.
RussianSensation
Elite Member
- Sep 5, 2003
- 19,458
- 765
- 126
I am not saying those results are bad, but they are not stellar. I don't disagree that NV is doing much better than AMD. My point is I copied those results from the 8 page CFO commentary by NV. I realize that comparing sequential GPU declines is not fully representative due to seasonality. Moreso, investors are probably disappointed by a relative low guidance for Q1 2013. Investors know that both NV and AMD need to find new growth opportunities outside of traditional dGPUs. It's not to say that that dGPU market won't grow, but investors nowadays are too concerned about the revolution/consumer trends towards faster growing markets such as smartphones and tablets. Both NV and AMD are under the same scrutiny, with NV easily superior financially, strategically and on the execution side.
Wishful thinking. Any time I say anything that people don't like in regard to NV, I get labelled an AMD focus group member or similar. Anytime I rip AMD's price/performance apart (like HD7970 launch for 3-4 months), no one says a word. If NV's cards were making me $ in bitcoins, my rig would be stacked with their parts. It's that simple. When the competitor offers GPUs that pay for themselves, I see no justification for paying $900 for a pair of 680s. At the current rate, two 7970s are making > $150 a month in bitcoins. I'll let you come up with reasons why someone who has cheap electricity should be spending money on slower cards that can't do this.
When HD7970 launched it delivered 20% faster performance than the previous fastest single GPU, GTX580.
GTX580 at the time cost $440 in retail vs. $550 for HD7970. ($550 - $440) / +20% advantage = $5.5 for every 1% increase in performance
HD7970GE right now costs $430 in retail vs. $900 for Titan. ($900 - $430) / + 60% advantage = $7.8 for every 1% increase in performance
Strictly from a price/performance point of view, the Titan is more overpriced relative to current flagship GPUs, compared to where HD7970 was relative to at that time the fastest single GPU (the 580).
Anyone who thought HD7970 was overpriced at launch would just the same think the Titan is overpriced. It's simple math. No double standards, no brand biases, just math.
Since the Titan is targeting the highest end PC enthusiasts, it would have sold out even at $1,000 anyway. If only 10,000 units of this card are made and it's a limited edition launch, NV has no incentive to lower the price. That doesn't change my point that people who ripped 7970's prices at launch but are now not doing the same for the Titan are hypocritical.
That's the thing. Some keep blaming AMD for letting NV raise prices. Company A does not allow company B to raise prices. Company B could take try to pass on higher prices to consumers. In the end the consumer votes with his/her wallet. If NV can sell cards at $1,000 without issues (GTX690), they will. If the consumer pays, they'll just keep those prices moving forward. That's simple capitalism, and little to do with AMD "allowing" NV to do this. The only person who ultimately determines what price NV can get away with is the consumer. If the consumers are fine now with flagship GPUs being $900-1000 instead of $500-650, then that's the new direction for the industry and we just have to accept that the ASP of GPUs is rising for both AMD and NV.
Last edited:
The only way nvidia will be able to get away with it is if AMD can't come close to matching performance. With the 40nm generation, we saw AMD stay competitive enough with nvidia's fastest that prices remained in check. Now we are looking at a situation where there is going to be a very large performance gap between amd's best and nvidia's best.
They beat The Street expectations for Q4 but... outlook Q1 2013 is not quite what was expected so that's why shares are going south. Nothing to worry about, that's Wall Street, shares fell when it was RUMORED Google will not use Tegra in the next Nexus 7. They need a few major publicized design wins for Tegra 4 and the damn LTE to impress the stock exchange. Those guys don't care about Titan and the fact NV is leader in desktop GPU, it's the "post-PC" era as they say and all that matters is mobile mobile mobile chips.
n0x1ous
Platinum Member
- Sep 9, 2010
- 2,574
- 252
- 126
Solid Post and good points. I think the bitcoin argument will be eventually irrelevant, but certainly a huge bonus for those that have cheap electricity and want to get into the virtual currency.
I think a possible reason for Nvidia getting away with the huge Titan price is that this going to be the first time since 2900XT vs G80 that Nvidia is going to have a massive lead again against AMD's best part. AMD stayed close or led at times with their dual GPU parts in 3870 x2, 4870x2, 5970, 6990, but with no (official) double up on this round from AMD, Titan and GTX 690 have a huge lead. I may be wrong, but I don't remember a huge fuss about GTX 690 price as compared to Titan. I think Titan is going to give the 690 a run for its money in many games due to the advantages of single GPU and in that case the rumored $900 falls right in line.
I can't afford to upgrade to a card that expensive, but I don't blame either company for the price. Nvidia has to answer to shareholders and will charge what they can to maximize their investment in GK110 and there is nothing wrong with that. Its not as if we are talking about some necessity of life. I don't blame AMD for not being competitive enough to make Titan cheaper because they don't have the resources to invest in a true competitor to Titan and the ATI team does a great job considering the rest of the company drags them and their resources down.
blastingcap
Diamond Member
- Sep 16, 2010
- 6,654
- 5
- 76
He always says that stuff about bitcoin but right now BTCs doubled in price in a short amount of time. It's not clear how sustainable these prices are. And he never mentions how ASICs have ALREADY been received by some people and will continue to be sold and shipped and turned on. Each minirig is worth dozens of 7970s. Therefore the difficulty of mining will skyrocket and you can't just assume that current profitability will stay the same. Prices might go up or down, but difficulty WILL rise, and that means profitability WILL go down. I'm not disputing that you can eke out some money during this transition period, but as ASICs pick up steam it's pretty clear that profitability of GPU mining will crash for anyone who doesn't get free/almost-free electricity.
Price/perf at the top end almost always sucks. And they HAVE charged almost $1000 before for a single GPU. Look at the old 8800 Ultra prices and adjust for inflation and you're getting a little uncomfortably close to $1k. Of course that was a bad buy on price/perf just like Titan will likely be a bad buy on price/perf (I say likely because I'm not sure what the real price will be... I'm assuming $900 based on rumors). That said, for some people (the 1%?) $900 is a drop in the bucket and they will buy anyway.
Last edited:
How can anybody not be impressed with nVidia? They have now nearly $3,8 billions in cash. Even if they have a negative earning of $200 millions every quarter they can survive the next 16 quarters or 4 years.
nVidia archived a record year of revenue and margin. They invest heavily into their future with a new campus, Grid, Tegra and Tesla.
And they started to give money back to their shareholders.
nVidia is here to stay for a long time.
nVidia archived a record year of revenue and margin. They invest heavily into their future with a new campus, Grid, Tegra and Tesla.
And they started to give money back to their shareholders.
nVidia is here to stay for a long time.
n0x1ous
Platinum Member
- Sep 9, 2010
- 2,574
- 252
- 126
Agree. They continue to thrive in a difficult environment. Great leadership from JHH has guided them through the storm.
Elfear
Diamond Member
- May 30, 2004
- 7,169
- 829
- 126
Look at Intel. We jump up and down and clap our hands when they increase performance by 10-15% every three years (sometimes less than that). People still go out and buy their processors even with very incremental increases. So while prices haven't skyrocketed, the performance we get is pretty lackluster. If AMD goes the way of the Dodo, we could be in the same situation with Nvidia.
n0x1ous
Platinum Member
- Sep 9, 2010
- 2,574
- 252
- 126
3 Titans - Obviously the extra $ doesnt matter to you (or this theoretical buyer)
Better scaling out of 3 GPU's than 4.
More memory bandwidth and VRAM for the obvious high resolution muli-monitor setup it would be driving.
Last edited:
RussianSensation
Elite Member
- Sep 5, 2003
- 19,458
- 765
- 126
Thanks
Will the Titan be limited to 10 thousands units or be in channels as a replacement for GTX690? If the Titan is a 1 shot wonder, then I don't think it'll matter for the market of dGPUs. Even if the Titan replaces a GTX690, and we consider it the new flagship, that implies HD7970GE/GTX680 are mid-range parts. That would effectively mean AMD/NV managed to raise prices on the rest of us gamers to $430-500 for mid-range parts. Both NV and AMD would be happy with that, don't you think? ^_^
You are right. They are mostly concerned about NV growing outside of the traditional GPU business, smartphones, tablets, etc. Is the Titan then just a limited run of failed/excess K20X chips that are better to sell off at $900 than throw them out?
True. When bitcoin ends, will we have to choose between $400-500 mid-range parts and $800-1,000 parts then? Will AMD and NV try to raise the ASP of GPUs to make up for the elimination of the <$100 dGPU market and slow growth of PCs? That's what I fear.
All good points. The consumer ultimately deems if the Titan is reasonably priced. Since NV has a very strong Professional solutions group where they already sell 500mm2 chips at $3-5K, the large die strategy seems to tie well into their overall graphics strategy. AMD doesn't have the luxury of making a 500-550mm2 chip for these markets and doing so only for the gaming market is prohibitively expensive. NV is servicing 2-3 distinct market segments with GK110 which justifies the expenditure due to the very high profitability they can net in the Tesla/Quadro segments.
Cash is considered a redundant asset by investors. What investors are looking for are growth opportunities for future cash flows and assess the risks and ability of management to achieve those cash flows by securing/winning those new opportunities. For investing money in a stock, they expect a commensurate rate of return based on the underlying risks associated with generating those future cash flows. Investors are not impressed if your business is strong but you have limited prospects for future growth and have a low dividend (same situation Intel has). Investors view Nvidia primarily as a GPU business (little growth opportunity) that's trying to expand into other markets with high growth opportunities (portable gaming, smartphones, tablets, etc.). Investors would not invest in primarily because of its new campus, cash position or its GPU business because those do not have strong growth prospects/ability to generate much higher future cash flows. A corporate building is just a real estate asset that depreciates in value unless you rent it out. If anything this is negative for NV's cash position in the short term because it implies allocating cash flows from the main business (R&D, SG&A) towards a real estate asset that in-and-of-itself does not generate cash flows. Additionally, they will have to capitalize the expense on the income statement for 3 quarters and then amortize the depreciation over the useful life of the asset. Investors are not impressed because they are not impressed by NV's ability to move away from post-PC era of GPUs as quickly as they would like. This is reflected in the guidance for Q1 2013 which doesn't show future growth that they expected:
"Nvidia’s revenue outlook for the next quarter fell short of expectations. The chipmaker estimates its revenue in the current quarter will be $940 million, not even close to $1.067 billion expected by analysts."
They also view NV's key competitors in the smartphone/tablet markets far ahead (Samsung, Qualcomm, etc.), which increases the risks/costs of barrier to entry for NV to secure design wins in those markets.
----
AMD has a similar problem regarding future growth opportunities, excepts its key business is actually dying. The risks are even greater for waiting until AMD finds new markets opportunities. Investors primarily view it as a dying CPU/APU/server business that's desperately trying to find new growth opportunities in other segments. Both NV and AMD are caught between trying to find new growth opportunities outside of their respective traditional business lines and both have not been very successful at doing so. NV has done better as a business overall than AMD which is why their stock hasn't been beat up to the same extent.
How many ASICs have shipped? 1-2? You've been saying that GPU mining will be unprofitable last fall, as far back as September at least. Back then people who listened to me went with HD7900 cards and paid for most of them by now. Right now the difficulty is so low relative to the price of coins, that even 1 month of mining on two 7900 cards would generate $100 in pure profits. Add to this that 7900 cards cost less than GTX600, and come with AAA games. All these factors together make them much cheaper overall. Price/performance is important to some people.
No one is advocating someone purchase HD7000 specifically for bitcoin mining. That's what you keep missing. Some people come on the forums asking what GPU to buy in the $300-800 range. No one here tells someone to go out and buy 10-20 HD7000 cards for mining. If you are already considering HD7950 x2 against GTX660Ti in SLI, it should be mentioned since it can save people $ on their purchase.
Last edited:
- Status
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 24K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
-
Facebook
Twitter