Content nec_000 - Page 6 - HWzone Forums
Skip to content
  • Create an account


registered user
  • Number of messages

  • Join

  • Recently visited

  • Days Won


Everything published by nec_000

  1. Good idea Yoav, ticking for us the processor models you wrote down, I wonder what happens in the said comparison. First 9400 vs. G3400: and now 9500 vs. G3400: Just so they do not say that the DB is screwed, because it does not seem logical to anyone that AMD controls (now) a lot of these processors, which usually house institutional computer systems, so I intentionally choose a processor that sells Certainly much more Intel than AMD in the private market, amateur computers in the field, model 7700k versus ryzen 1700 - hits of 2017. That we will see it all together, that we will understand that numbers are correct and there are no games here I think. The numbers are of course going down in both, because these processors are no longer sold new, so what is in the field is getting smaller and smaller than the general cake as newer processors enter the market. Therefore, we are seeing a gradual and slow decline in these two models, which corresponds to reality and is that their share as a percentage of the total number of processors in the world is declining. Want to sample more processors to be convinced of the trend of the last year (two years), about the shift that has taken place in the processor market in the field of desktops? Or is the trend already quite clear to us in light of the data we have collected? Of course, all the factors were gathered and mentioned: including from retailers in Europe (mindfactory) and USA (Amazon), continued research institutes that do their work (for market analysis as a recommendation to an investor in the capital market), and ended (in my opinion) with the best data and DB engines. There is a type in the field, a screen of all the processors in the world at that moment.
  2. In economics it goes like this: if there is a shortage in the market compared to the demand, and therefore the price rises, this means that profits rise, which in turn created an incentive for players in the field to increase output. To sell more and earn more. To say it will take two years to balance? Seems to me a little too long. Unless it is an infrastructure problem, some kind of barrier is interfering. In this case of silicone chips, the raw materials cost zero and are not limiting. Manpower is not lacking either. Mechanization infrastructure takes time to organize and implement on the production floor and this is probably the biggest limitation at the moment. I estimate / hope that the gap between supply and demand will be balanced within a period of less than two years, hopefully within a year.
  3. For this lanzar we need two things: 1. The first is that you will already get some 6900xt model that is very difficult to catch as we noticed, then we will be smarter in terms of frequencies it is able to perform at a higher ceiling light = 3000mhz instead of just 2800. And how hard to catch forum members need Understand: Unfortunately even the 6900xt red devil you ordered for some reason canceled my understanding in newegg his order and fucked us to the end as they say. Now we have to wait for it to be back in stock and you will know if they will not do this exercise for us again? 2. The second is the one that converges to a customized Bios experiment on all its risks. The taichi card we saw contains two bios so the game of burning one of them does not endanger the card, because if it turns out to be damaged, you can move the card from the other bios that is still working and run the other bios again. The problem with phantom cards is that they do not have dual bios, and then the danger of overriding the single bios using custom, is that if it is damaged, God forbid, the card has gone to waste, and it is 4K chess that is on the table. a little scary. I therefore recommend that you try to grab one taichi card, and the bios experiments from Custom will be conducted on it. So as not to take the risk of destroying such an expensive card. One way or another, there is no doubt that the new generation of cards launched in the last quarter (Empir and RDNA2), opened a whole new benchmark of performance and brought the 4K resolution for the first time, to be the resolution that is practical and real to the consumer. The previous generations claimed to be 4K, whoever remembers, claimed 1080ti and then claimed 2080ti, but in practice it turned out to be a ball that is only suitable for 1440P if you want a reasonable FPS rate. This upgrade that brings with it the new generation and is a practical ability to run nicely on 4K, is what will lead to a change of trend in the screen market, when 4K screens will start selling in solid quantities and increase the penetration share of this resolution in the market, which has not happened so far.
  4. Regarding sales data from etailer suppliers, it can be assumed that these are indeed private consumers who buy a computer at home, such as from the German mindfactory supplier or the American amazon, and not in the institutional market, which purchases from other suppliers in the distribution chain. But ... research data from institutes that provide background data in favor of preparing a prospectus for investments in the capital market, and the data from the most recent database I provided us - in processor comparisons with processor blocking leading segments, such as 10900 versus 5800 and so on ... these are no longer data from Any stores or marketers, but data that is extracted from all the computers that go on the road in the world and connect to the Internet. Are they the more accurate and comprehensive picture that we might want to look at? So I would run a query for us on the Database, specifically on the processor models that usually go to Lenovo Del and HP. What might be a different picture? After all, we will see more sales of institutional market processors. So throw back what you think are processor models. The most common in these computers, and I will run a query. It's interesting. I'll assume in advance that these will be quad-core processors and frequency-locked cores. In my opinion, on the part of AMD, these will be APU processors, since an office computer also needs a video card. The quad-core commonly used in branded computers, as opposed to AMD's corresponding APU. What models to put in front of what? Help me
  5. ** I can test each model of processor and compare it to another processor. Ask what I will tick in the DB engine
  6. The later arrival of the 10400f model (at a slightly cheaper price), failed to reverse the trend, but at least lowered the ratio from 9 times in favor of the Risen 3600, to only 5. times the difference: Yahya the difference:
  7. I was intrigued by what has gone on here in the last year, comparing the two Kings, 10400 versus 3600, the popular 6-core princes, because they talked about these two processors head to head the most on the net. This was the talk of the day, who is the ruler of the mainstream category, and the truth is, I did not know it was the actual gap, inconceivable. In fact the 3600 processor we knew was one of the best selling and this is not surprising, turns out to be in the first place in the world with a market share of over 4%, a record that was unparalleled I think very much Sandy Bridge. It sold 9 times more than its direct rival 10400 processor:
  8. More interesting business statistics, in direct comparison of the segment which the consumers of the premium desktop buy for a home performance computer, and can not be confused here. Comparing Risen's new 8 cores = 5800X against the 10 most powerful cores Intel 10900K has, and accompanied by dates that show us what happens from the moment the Risen 5000 launches and how Intel's sales graph slows as Risen's sales graph soars, at least until full of it on the shelf. (Probably in the December area): And here's a comparison of quantities against the Risen 5900X, more or less the same. Also for sure see there is a sales halt in the last month of December. Undoubtedly a supply side problem, which does not allow customers to get the products near: here is a super fascinating comparison of what happens in the center of the table where there are more sales, the $ 300 MSRP price segment, in the tough competition of 10600K versus 5600X, the six cores representing the sweet spot today. Here too and even more that the Risen is sold 3 times more than the Intel in question and again, it is impossible to confuse the market shares and say it is just some statistical sample of passmark and therefore take a limited warranty. It's everywhere and every nook and cranny that is measured at every business economic data provider that is available:
  9. Amazon USA also provides product sales data by category, and a quick tour of it shows: that AMD processors control by hand Level in sales volumes. Not half and half, but similar to Germany, great control of sales of AMD processors. Its 1-6 places, 7th place one Intel processor and after that goes on again 8-9 AMD, instead of 10 again one Intel processor, and so on and so forth. Clicking on the link will show viewers an up-to-date picture. I'm taking a second for us to see what the page currently shows in the first places at least and the models:
  10. Beauty, I found the most up-to-date, for the month, including closing data for December 2020 from a week ago. ** This time it includes the launch of Risen 5000 and also the launch of Intel 10th generation, ie more current than ever. Although there has been a recovery on the Intel side and total quantitative sales have increased (see in the graph representing Intel a nice increase in nominal numbers), in contrast you can see that the increase on the AMD side is even greater and the jump from the launch of Risen 5000. Very interesting and instructive:
  11. A quick browse through the web to see what's happening in the market share reported by the German etailer mindfactory, updated for April 2020 (before the launch of Risen 5000) paints the following picture: * I will look for more updated if I can - stay tuned
  12. In a test I did last month, you casually asked on a TMS of the kind I ask when hugging friends once in a while ... when it's popping up to buy something, and it's "What's the attitude today in selling Intel processors versus Risen with you?" I was seated "selling about half and half." And the same question in 2017 about six months after the launch of the first-generation Risen, was answered "a quarter versus three quarters." The so-called peddler statistics from the market, in one crazy little store in Israel. As far as it means anything to us. And this is Israel, yes, not the United States or Germany here. We did not ask to present what is happening at the large mindfactory retailer in Germany that has long been pointing out these findings to the community, but rather asked to see what is happening in one typical store in the Holy Land. Given that these findings, of mindfactory, of all kinds of economic institute research for the investment market, and findings from a local street shop here converge to the same / similar picture, I estimate that passmark's statistics are very much in line with the mood in the global desktop market today. The private ones (assuming that office computers are not included in the sample for the purpose of the subject).
  13. Following on from the issue of the differences in the image seen between AMD and Video and the need to investigate the issue, a procedure that will no doubt take some time, I have managed to gather the following material in the meantime: the image sharpening that AMD introduced as part of the image enhancement package it introduced last year, along with RDNA1 launch Known as the 5700xt, then also Nodia has already managed to complete its driver and has the equivalent. Maybe even the same. Here is a first overview of the issue in hardware unboxed. Just note that since the review came out before DLSS2 then one should take accordingly the reviewer's remarks, which are a bit old in today's terms. Second, in the reviewer's update, AMD has also already updated the driver and is now working on both DX11 and other cards and not just RDNA, and it is also possible to calibrate it on a game basis. That is, each of the manufacturers learns from the other and copies what is good in the other and it is beautiful like this: here is a demonstration that someone did with the function is activated and aggressively configured for maximum power, so that we can get an impression of what it does more easily. The truth is, definitely interesting and worth checking out: and here too a demo with and without a picture next to a picture:
  14. Meanwhile the economic research departments are issuing interesting data, in which quietly AMD has penetrated within 3 years with epyc processors to over 10% market share. One estimate speaks of 12%. The data is correct for November 2020, and unlike the desktop segment, these do not yet include AMD's move to zen3 - based epyc processors, but only the previous generation zen2. There are estimates that in 2021 using epyc zen3 AMD is expected to cross the 15% market share threshold in the data centers. ** At the same time there is also a big increase that ARM processors are starting to accumulate in this market. It is worth following the development as stock investors. Strong nod in this picture.
  15. Slightly disappointing news, the launch of the Navi22 known as the RX6700 under the market name is postponed to the end of March (planned for January). The reason: Lack of allocation of 7nm silicone at the manufacturer TSMC. After all, they are already not in demand, when allocation is also required for Risen 5000 processors, chips for consoles (from Microsoft and Sony), RDNA2 cards from the 6800 series, and epyc processors for servers .... It seems at the moment, the situation is not particularly optimistic. And if so, then they will make a preference - allocate silicone to what is most profitable in the product line, while giving up the rest.
  16. חזרנו לימים העליזים של Athlon מול פנטיום של אינטל
  17. You can definitely believe it. In the lanzar the second (better) phantom scratches 2800 from below. We saw about 2760-2770. So if the taichi is a bit better than it (hypothesis only as of this moment) then it can be assumed that it squeezes out the last 2 percent. As soon as custom-made bios come out for this product, which releases the 320w detector, I estimate it will open up some more juice here that can be milked from. Since the card has a dual BIOS, the fear of being trampled on by an external BIOS drops off the table. Max BIOS not suitable? Rather, turn on the card from the second BIOS and override the damaged one once again.
  18. You can see that the platform with which he performs the Intel 10900 test limits the cards and does not let them flourish. Especially the RDNA2. Especially in 1080P but also in 1440P. Another interesting and not surprising thing is that once it moves to a more efficient Linux operating system, the business starts to free up. Which only shows us the inefficiency / inefficiency of the windows operating system, and it also does not rework us anything for the truth. I liked that it gives a close-up look and high photo quality on the product, which allows us to get a good impression at all of the size and technical quality of the card. Overall did a nice job that complements and helps the respectable work he did for us lanzar. Further provided an impression of the noise level and explained that when it's inside the case, you hear nothing. Ie silent product practically. Which is the modern trend of tickets in recent years by the way, and nothing out of the ordinary.
  19. Lanzar, try to order a Taichi card, let's compare and see how it compares to the phantom. Do you know maybe a little better? Still it is expensive at some 100-200 NIS ....
  20. We asked to keep up with the improvement obtained in the 6800XT speed, ie how much performance we get - in exchange for a few percent speed in the core operating frequency. To do this, Lanzar ran two measurements to describe the noise, one at the stock frequency as AMD originally released and marketed the product, and the other at full speed, the one that can be reached in the given test, which in the case turned out to be 2710mhz (because this is the full frequency reached in the individual case Of the benz chosen for the subject). He used a tool called the unigine superposition. This is an excellent tool that is easy to use for apartment experiments and it neutralizes the layer of the processor from the equation, since it does not load the processor with anything. It tends to be a graphic test to see efficiency in this layer that we are trying to rub against. The following are the findings: the frequency was improved by 23%, for which a measurable performance improvement of 21% was obtained (88.25FPS versus 72.91FPS). From what we examined during the past week it is more or less what we also saw, that when the CPU / platform is not the interfering factor, efficiency coalesces to 90% improved performance for frequency increase. Which is impressive and quite linear. It is clear to us that this is possible due to the fact that VRAM does not interfere with the process or restricts, since speeding up the core directly also speeds up the cache in a 1-to-1 ratio, the one in front of which all the work is actually done. Therefore 90% utilization is an expected and not surprising figure in any way, because of the architecture we are dealing with:
  21. As it seems, coming 2021 marks for the first time the transition to core touchdowns, as the new standard in the new and heavy titles on the market. It is possible that i7 quad-core i7700s (from Intel generation icore first to seventh 10400) that have active HT help in some way compensate, but the question is how much and for how long one can survive with it. In my estimation on borrowed time only. So right now for those who need a new computer, the recommendation would be to make the effort whenever possible, and purchase a 700f min processor that is hexagonal with HT for around 10100 NIS, and not to compromise on a 300f quad-core for around 3 NIS, to be what is called with the required power balance For the next XNUMX years. As it seems, we have reached this moment with respect.
  22. Undoubtedly, in a business field where there are only two contenders, they cannot strategically afford to lag behind too long. When one does something that provides a nice leap forward for her, the other immediately comes and copies the patent. We have seen this for almost all the years. We are currently faced with two performance technologies that constitute founding events in this industry, including cache memory and multi-core on a unified PCB. I estimate that each of the actresses (the only ones there) will adopt these two techniques no more than one year after the other. Because whoever does not adopt loses here a significant advantage that it gives to the other competitor, and therefore can not afford it.
  23. In light of the finding that in the 6800xt model, AMD deliberately limits the ability to speed up to 2800mhz in the slider so that the working frequency can no longer be increased, and there are cards that do the specified maximum and are not excited about it (for evidence we just see above), and in light of the 6900xt model the slider allows Up to 3000mhz, we have a preliminary estimate in our hearts at this point, that AMD is doing so on purpose: in order to give the 6900xt versions the ability to work at a higher frequency in order to keep it a little further than the 6800xt, to differentiate them into sufficiently separate performance levels. So when women get our hands on a quality 6900xt version we will bring this theory to the test and report to you neatly.
  24. In the meantime, lanzar started measuring and playing with the second asrock phantom, only to find that it is a better card than the previous one with which we did all the measurements in this thread is a better binned card. Attached below is the video he transmits. This card makes a stable 2750-2770Mhz Also in the video answering the questions of the participants in the thread, the working temp they see, it's in default fan settings on auto. Ask the caveman mostly (and one more friend), so we got feedback with measurement True. And yet the temperature is between 56-60 degrees typical. It's just very impressive the air cooling of these asrock cards, the size and weight do their thing, no tricks. Which implies to us that there is really no need to think or squint for water cooling. Totally unnecessary. ** Like Yes lanzar checked for the scientific purpose of the matter and as a comparative reference, what happens when he raises the fans to 100%, only to find that in this model of asrock, the improvement in lowering the temp and not worth it at all. So it is also unnecessary to do so and better leave the fans on auto - it's most true like this.
  • Create new ...

At the top of the news:

new on the site

Now it's official: the comeback of the GeForce RTX 2060

Now it's official: the comeback of the GeForce RTX 2060

The intermediate model from the Turing generation adopts the number of active units in its refreshed super version, equipped with double the volume of graphic memory - and goes out to provide a solution to some of the constant demand for more and more video cards