9700X & 9600X Instant Reactions (Die Shrink Telegrams)
Added 2024-08-08 07:04:41 +0000 UTCWhat else are we going to talk about? Write-in below with your immediate thoughts on the 9700X, 9600X, and Zen 5 in general! Obviously, we'll welcome all AMD & Intel questions here as well. Just be respectful, concise, and use good grammar to be considered! You have ~16 hours to submit!
https://www.anandtech.com/show/21493/the-amd-ryzen-7-9700x-and-ryzen-5-9600x-review
https://www.phoronix.com/review/ryzen-9600x-9700x
https://youtu.be/JZuV35LgjxU?si=_tGzbeJ82Z96a--9
https://www.tomshardware.com/pc-components/cpus/amd-ryzen-5-9600x-cpu-review/2
https://youtu.be/OF_bMt9fVm0?si=UoPsDOuMjOXf10nJ
https://youtu.be/DWYFfzHFtM8?si=v1EQbiKfjWkmU0Iy
https://youtu.be/DWYFfzHFtM8?si=76eigaD50x_B_S7L
https://youtu.be/jPJ0Khw3kIc?si=P7WGlejQ-jyl2CIG
https://www.techpowerup.com/review/amd-ryzen-7-7700x/24.html
https://www.youtube.com/live/MxOL56hjXW4?si=RV4R7sWXgxNRxfIW
Comments
Clearly Zen5 cores are focused on server performance and power savings. I'm impressed but gamers who love desktop heaters won't be happy.
Kiln God
2024-08-09 13:42:25 +0000 UTCI'm a bit disappointed with the gaming performance of the 9700X compared to the 7700X. I was hoping at least a 10% improvement in performance at stock settings, but the uplift I was hoping for was close when compared to the non-X 7700 (same 65w tdp). With PBO and other tuning, the 9700X shows decent improvement, but the extra power doesn't seem to enhance the gaming performance. Could it be that AMD opted for a low TDP because gaming performance doesn't differ much even with the same power consumption?
Leifrobert
2024-08-09 12:03:39 +0000 UTCHi Tom and Dan, to me it looks like Zen 5 has the primary focus on Server/Workstation/Mobile as it shows impressive gains in efficiency and compute for Single/Multithreading according to phoronix. Do you think Gaming was an afterthought and will that change over time with optimizations? You said Zen 4 is aging better and catching up to Zen 3 X3D due to the newer architecture and featues.
KingKoro
2024-08-08 22:33:07 +0000 UTCI hope AVX512 dies a painful death, and that AMD starts fixing real problems instead of trying to provide magic instructions to then create benchmarks that they can look good on. Because absolutely nobody cares outside of benchmarks. The same is largely true of AVX512 now - and in the future. Yes, you can find things that care. No, those things don't sell machines in the big picture. I hope AMD gets back to basics: gets their design process working again, and concentrate more on regular code that isn't HPC or some other pointless special case. I want my power limits to be reached with regular integer code, not with some AVX512 power virus that takes away top frequency (because people ended up using it for memcpy!) and takes away cores (because those useless garbage units take up space). Stop with the special-case garbage, and make all the core common stuff that everybody cares about run as well as you humanly can. Then do a FPU that is barely good enough on the side, and people will be happy. AVX2 is much more than enough.
【ｓｃｅｎｔｗａｖｅ】
2024-08-08 22:22:05 +0000 UTC“Skylake-level uplift/efficiency gains” moment. Not sure how a generational gain like that can ever be satisfyingly marketed to consumers. Skylake-SP-style technical strategy/HPC market-focus too, with consumer stuff now having both less avx and less cache. The microarchitecture/hard-design is diverging in the same way.
【ｓｃｅｎｔｗａｖｅ】
2024-08-08 22:16:03 +0000 UTCCould it be that AMD went for efficiency over performance for zen5 because of the bad press Intel's been getting with their power draw for the past few years? and especially now with the degradation issues. Or could it be to get some wind out of Intel's sails in case Arrow/Lunar lake is a bigger efficiency uplift than AMD expected? I'm all for being at the peak of the efficiency curve at stock settings, but when there's barely a performance uplift over last gen it feels kind of pointless.
Arknia08
2024-08-08 21:34:34 +0000 UTCI think zen5 for desktop was kind of an afterthought for AMD. When given the efficiency, small die and focus for professional application performance, it seems to me that zen5 was basically entirely designed for mobility/server. I am curious how well these chips undervolt. Given i can already undervolt my 7700x to where it almost meets the 9700x in efficiency anyways. (At -125mV it uses about 80-95W vs 145W stock) I am glad overclocking is back though, and these cpu's aren't being redlined out of the box anymore. Overall I currently think of zen4 to zen5 as a skylake to kabylake situation. Same performance, slightly more efficient, higher price. This feels like stagnation the second competition is removed. Let's just hope this isn't history repeating and we won't be looking back at "AMD's decade of 16-cores" in 2029
Arknia08
2024-08-08 21:03:45 +0000 UTCWill amd take Intels approach and put full zen5 desktop chips on laptops? They could take the crown!!!
Manordown
2024-08-08 20:15:20 +0000 UTCBefore or after the launch of Zen 4, I’ve not seen any indication that AMD went to any effort to promote AVX-512 to developers or assist them in any way with implementing it. The only application or benchmark AMD seems to have influenced is y-cruncher. At this rate, I see little changing by the time Zen 6 comes out.
Chris Rijk
2024-08-08 17:14:15 +0000 UTCDo you think there's a disconnect in the marketing department at AMD between features, performance and price/value? For starters, we know perf per watt and AVX 512 perf are features that AMD are pushing (and this may primarily benefit server products for Zen 5) but the consumer price is high compared to the out-of-the box performance compared to Zen 4 products already on the market. But if those features are ignored, or pushed to the side, then the value of the consumer Zen 5 products becomes greater. Are AMD marketing unable to differentiate between different market segments? We already suspect that they're pretty incompetent (as we've seen over the recent years on both Ryzen and Radeon products) but is this a case of them being unable to identify a particular message to focus on when designing the product for retail?
James Prendergast
2024-08-08 16:21:49 +0000 UTC1. Is the 9000 series the moment where AM5/DDR5 makes sense for a platform upgrade from AM4/DD4? -- I think so, the 9700x & 9600x both provide a mainstream price with class leading thermals, class leading power use, and class leading overclocking headroom. 2. If you were an average users, light gamer what other CPU offers a better balance of price/performance/power? 3. Based on the branch prediction and memory optimization that Wendel saw, do you think we will see the higher end parts, and x3d parts widen the gap? I just dont see the disappointment or agree with the "Flop" perspective that HWUB offers up. When I look at the data, the single core performance is generally top of the charts, 1% lows are top of the class, and the LOWEST end part offers reasonable gaming at 4k with only 6 cores. The tests and charts often show AMD matching or beating Intel with 1/2 the cores at 1/3 the power. All that is before even talking about the fact that the only "safe" generation of Intel to buy right now is 12th gen so putting 13&14 generation parts that no one can recommend on the charts is not even a good comparison.
infinitevalence
2024-08-08 15:31:56 +0000 UTCWith the Zen 4 launch, AMD got burned at the stake for pushing it hard out of the box and everyone thought 95C was the end of the world. They also praised how well it still ran when running in "Eco Mode" with many preferring that configuration. Now AMD launches Zen 5, makes Eco Mode the default and you can still push it as hard as Zen 4 with PBO and everyone is still losing their minds. AMD really fumbled the messaging on this one. They should've kept the Zen 4 TDP and said "hey check out eco mode too, it matches last gen with a 40% power reduction". Or said the default power is designed for efficiency and enthusiasts need to turn on PBO. It's really just a messaging thing when it easily can do both with a flick of a switch. A stagnation in value? Sure. A technical stagnation? No.
shredbird
2024-08-08 15:07:41 +0000 UTCDo you think it's possible that out of an abundance of caution AMD kept the 9700x power and voltage down since it is 2B more transistors than 7700x in similar amount of space given what is happening with Intel right now?
Dark Helmet
2024-08-08 14:56:01 +0000 UTCI mean I personally bought a 7700x with free ram at microcenter. I don't regret it. I'm playing at higher graphics so I'm not CPU bottlenecked.
B. Fish
2024-08-08 14:45:39 +0000 UTCThe zen 5 reviews have me really concerned for laptop. Strix was somewhat underwhelming on the CPU side vs hawk point, do you think by the halo launch they'll have everything worked out?
B. Fish
2024-08-08 14:44:09 +0000 UTCIs ddr5 6400 the top reliable memory speed for Zen 5? I was hoping to see dddr 7200 somewhere
Dig Wiggler
2024-08-08 13:45:57 +0000 UTCHi Tom & Dan, any evidence to suggest AMD realizes they fumbled this launch? Would like to think they are learning the lessons at least or shall this be a trend that continues to happen?
Gian Violi
2024-08-08 13:39:12 +0000 UTCDo AMD X3D variants make other 8 and 12 core models "bad purchases"? With Zen 5, I fail to see any reason for someone to buy the 9700x or 9900x. If they need multicore, buy the 16 core, while for budget buyers probably 7600X or 7500F. In gaming, it seems like you wait for the 9800X3D or buy the 7800X3D at $350 like you have been saying for months. I don’t see why AMD continues to sell the 8 core variants without X3D, even at rock bottom sale the 7700X only really was $50-75 cheaper than the X3D at microcenter.
Glass
2024-08-08 13:36:22 +0000 UTCHi Tom and Dan, I would like to ask 1. ur opinions on the pricing? IIRC Tom used to say AMD will be aggressive at pricing for zen5 but it doesn't seems to be the case especially 9700X vs 7700X barely have any performance gain for whatever reasons 2. Does AMD have any people who have technical background or at least know what products are they making in the marketing team? It seems to me that they f'd up the GPU side and now also f'd the CPU side after they have been somehow consistent on CPU for at least 5 yrs. 3. Who or which department in AMD should take the biggest blame of this epic failed launch of zen5? And BTW, my ID is pronounced like the "ming" from "swimming" but closer to "main" :D
MinG3825
2024-08-08 13:28:45 +0000 UTCHey Tom and Dan, do you think this AMD launch is the small crack Intel needed for their Arrow Lake launch? If Intel can come up with a robust RMA system like how you have suggested in multiple Broken Silicones, demonstrate a 10% single core IPC increase with good efficiency, then a 20-25% IPC increase with a higher power profile (hopefully not 230w+), and pretty much give reviewers permission to break their cpus to see how long they can last. Do you think this can keep Intel in the game?
Kinihun25
2024-08-08 13:17:12 +0000 UTCDo you believe AMD saw the Intel drama and decided to throw out lower quality chips under a higher designation? If you cannot get more capacity to take advantage why not get higher prices? Why not just release the true 9700X as a 9750X later. This would help you get rid of old stock and lower quality inventory at better prices before the real launch?
Bryan P.
2024-08-08 12:57:29 +0000 UTCIn their presentations, AMD did show specific results for the 9700X and while some can be replicated, some are way off. Is this another RDNA 3 situation where the performance is lost forever? And if so, what caused it?
Chris Rijk
2024-08-08 12:47:55 +0000 UTCDo you think Robert Hallock's departure from AMD played a part in the problems with marketing Zen 5? If so, what do you think AMD needs from a Director of Technical Marketing to prevent similar screwups in the future?
Cleansweep
2024-08-08 12:06:17 +0000 UTCThe biggest disappointment to me is the lack of improvements in memory speeds. That seems to be holding it back in a lot of areas, from what I could gather from the few reviews I watched. I don't plan on upgrading anytime soon, very happy with 5800x3D, so even if Zen5 was cheaper than Zen4 street pricing, I wouldn't be in the market. I just hope that AMD isn't starting to stagnate now, since I'll be very keenly looking at Zen6 and after.
Trogdor
2024-08-08 11:38:34 +0000 UTCDo you think that unless AMD changes strategies with Zen 5 positioning this is giving Intel a chance to potentially make Arrow Lake look a lot more favourable, at least on the performance side?
CompressedAIBlocks
2024-08-08 10:54:32 +0000 UTCDo you think people have taken for granted the out of box performance AMD has given us with Zen 2-4? They're really reserved here.
CompressedAIBlocks
2024-08-08 10:45:49 +0000 UTCTom I think you've shed some good light on Zen 5 in your analysis video. I do think it's a blunder from AMD to expect (want?) users to tweak their hardware to get performance that they left on the table. If this came out with the performance of say debaurer's review and people had to tweak for another 5-10% thats one thing but the reality is most users, even PC builders, still don't tweak besides maybe if there is a simple on/off switch like Eco mode. This could have walked all over Intel because of their bad image right now, however now AMD is sharing some bad PR. As an Intel users I was tossing up replacing Raptor Lake with Zen 5, now it's a wait and see at best from me.
CompressedAIBlocks
2024-08-08 10:43:03 +0000 UTCBetween the 9600x/9700x having pretty awful performance in gaming and Strix having "quirks" of very high cross CCX latency, so far Zen 5 seems like the most troubled of the Zens... Do they need Jim Keller back?
qhfreddy
2024-08-08 10:27:24 +0000 UTCWhen will the mythical DDR5 8000 AMD had on slides happen?
qhfreddy
2024-08-08 10:25:48 +0000 UTCWhy does everyone compare the 9700x (65w) with the 7700x (105w) when the 7700 (65w) is right there?
qhfreddy
2024-08-08 10:25:24 +0000 UTCI think zen5 is in a big part a stepping stone to zen6. With zen6 we can expect more cores so we need more efficient cores and overall there are lot of architectural improvements that will be fully realized once software is better optimized for zen5.
QuickJumper
2024-08-08 10:13:57 +0000 UTCI'm hoping AMD will deliver on the promise of supporting higher DDR5 memory speeds while still remaining stable, at the very least with 2 modules (48-64 GB of system memory).
Søren Klintrup
2024-08-08 09:47:36 +0000 UTCDon't care, am waiting on Zen 5 X3D.
Wasmachineman_NL
2024-08-08 09:29:44 +0000 UTCYeah, it seems for development work on Linux it's amazing. I am waiting for more people to start using. Then we will know.
Manolis Ragkousis
2024-08-08 09:21:35 +0000 UTCThe Zen 5 core itself looks good, especially as a base to improve upon with Zen 6 (with new packaging and hopefully a redesigned backend). The workstation/productivity workload improvements are very solid, but the gaming uplift is disappointing. Curve shaper is a great tool and I think enthusiasts will have a lot of fun extracting additional performance out of that and memory tuning. AMD really should have put more effort into the base profile though as the average user will not go through these steps and I think general perception as a result is that Zen 5 is a flop
GrandDemand
2024-08-08 09:19:29 +0000 UTCInitially I was a bit put off by these. However after seeing more posts, it seems that the messaging got botched. I'd like to see them retested with improved memory settings and see what is different.
Thalo215
2024-08-08 08:46:27 +0000 UTCShould this have been called something other than zen 5? It seems like a different beast (medusa, phoenix...) than zen 4. Maybe some alluding to ryzen like rye bread architecture or brazen architecture or frozen architecture. Zen +1 seems odd at the moment.
Nicholas Buckner
2024-08-08 08:43:13 +0000 UTCAll the reviews I read and watched today had broadly consistent results with each other, but Phoronix seemed to cast the widest net productivity-wise. About 5% in code compile, but the improved Python and node.js performance is something developers really care about. And the browser performance improvements are relevant across a huge range of workflows these days. Wendell from level1 also reports good ECC compatibility this gen on a number of motherboards. I could see this lineup as a solid contender for productivity. But I *really* don't like what I'm hearing from GamersNexus about stability issues and poor quality control. And a discounted 7950X in eco mode also seems like solid contender with many of the same efficiency benefits. We'll see how the 9950X turns out. What's interesting is how much better AMD's competitive standing is on laptop with this generation. Efficiency seems like Zen 5's strong point, but it doesn't seem like it scales well at higher power targets in these desktop socketed CPUs. Just shows where AMD's head is at this gen. Also, when a company touts its fancy new branch predictor, but SMT scaling is "really good," (per level1) I worry. High SMT scaling can be a consequence of low 1t IPC due to poor core utilization. This can leave a lot of room in the core for the second thread to operate. And branch prediction gone wrong could potentially cause low 1t core utilization, which might also explain why the core would refuse to pull more power than stock even with PBO in some workloads. The former Intel engineer Francois Pidnoel descussed this property of SMT in a video a few years ago. Whether it applies to this CPU is total baseless speculation at this point, but I think it's an interesting theory.
Max Eliaser
2024-08-08 08:08:37 +0000 UTCWhile you recommend waiting another month or two, hasn't Zen 5 already been severely delayed? "AMD ZEN 4 to ZEN 6 Leak: Transforming into a Premium Brand" suggested Zen 5 would come out 11-15 months after Zen 4. Zen 4 came out in Sept of ... 2022, about 22 months ago.
Nicholas Buckner
2024-08-08 07:51:33 +0000 UTCI actually think zen 5 looks really promising. Slightly better performance on much lower power consumption. I think the 9950x will be a monster if it has a 125 TDP
Manolis Ragkousis
2024-08-08 07:48:46 +0000 UTCWhat do results in non-gaming workloads (for example Phoronix showing 9700X with 15% gain on 7700X at 74% average power) show for Turin, and Turin's performance against upcoming Granite Rapids?
KarbinCry
2024-08-08 07:39:31 +0000 UTCWith Zen5 being exceptionally memory-starved, this makes the need for a double-X3D 9950X all that more juicy a prospect. But it seems like AMD is disinclined.
Nicholas Buckner
2024-08-08 07:38:38 +0000 UTCDo you think Ryzen 9000 will have FineWine factor? It is a huge architectural change, applications simply are not optimized to use the wide core. And scheduling will likely have to be adjusted to exploit Zen5's focus on better SMT.
KarbinCry
2024-08-08 07:35:57 +0000 UTCPerhaps the underpowering of the 9700X may, unintentionally, make Zen 5 more attractive as the CPU architecture for handhelds or thin-and-lights? With the lower power consumption still providing acceptable performance, maybe this is a silver lining. Doesn't change the launch being at best confusing, and at worst, making Zen 5 look like it's a nothingburger.
Woody Chang
2024-08-08 07:33:47 +0000 UTCGiven how 5800X3D did very well against non-Vcache Ryzen 7000 in gaming, why are people surprised that non-Vcache Ryzen 9000 is behind Ryzen 7000X3D in gaming, when it lacks advantages like optimization and memory bandwidth that 7000 non-X3D had over 5800X3D?
KarbinCry
2024-08-08 07:33:32 +0000 UTCI guess more than one 9600X arriving broken in reviewers hands absolves them of the whole "cherry-picking" accusation.
Nicholas Buckner
2024-08-08 07:32:00 +0000 UTCDo you think gaming is a "solved" workload on the CPU side?
KarbinCry
2024-08-08 07:31:37 +0000 UTCZen 4 had a slow meandering re-calculating of price with MicroCenter, via bundles with--at the time--expensive DDR5. I expect a slow meandering re-calculating of firmware and settings. Hopefully, they can get working a different performance profile for wildly different TLPs. I don't think AMD would go that far: AMD doesn't want to appear to have anything similar to Intel's "Baseline" "Performance" "Extreme" non-sense, almost to their own detriment. What makes reviewers and owners hate them is the lack of clarity. There was nothing wrong with "Eco Mode", or Eco mode out of the box. Call one "Gamer mode", and one "Cruising Mode". Hell, TDP170 Mode works. AMD appears to have over-learned this lesson.
Nicholas Buckner
2024-08-08 07:30:16 +0000 UTCI am looking to build a Linux workstation to eventually replace my 3900X. With the leaked 15% IPC increase, I was expecting a 0-40% spread across tests. Early benchmarks seemed consistent with that, so the HWUB review was stunning. The tremendous Phoronix and Level1 reviews put me back on the course for the 9900X or 9950X, however. If the R9s review well next week, I plan to pay the early adopter tax for a Zen5. My concern is that if the 9700X is memory limited, this may be a bigger problem on the 9950X.
Andrew Hansford
2024-08-08 07:27:47 +0000 UTCIt's almost as if AMD did the polar opposite of what Intel does during their launches. Instead of overclocking their CPUs really hard out of the box to the point of instability like with Intel, AMD undervolted the Ryzen 7 9700X so low that its out of box performance is sometimes weaker than it's predecessor. I just hope this doesn't take the wind out of the sails of Zen 5, because it seems like the architecture has a lot of potential with its power efficiency and IPC increases. Who knows, maybe this was all a ploy to get people to take overclocking serious again.
kjm015
2024-08-08 07:19:24 +0000 UTCIt seems many reviewers correctly pointed out, at it price point it doesn't make sense. However, I would point out with the absence of Intel as competition, its price point makes perfect sense. Another aspect many reviews overlooked, possibly due to not being their specialty, was its fornominal small factor performance due to low power consumption.
Baza's Sami Note II
2024-08-08 07:18:59 +0000 UTCOne of the best selling features of Zen 4 was its performance per watt. Places with high costs for electricity, like Europe, ate up Zen 4 due to this metric. They left performance on the table for PBO to claim back later. AMD appears to have over-learned this lesson.
Nicholas Buckner
2024-08-08 07:18:12 +0000 UTCThey‘ve really fumbled that one, huh? Anyway, I was just wondering why we don’t see any X3D chips on mobile. Especially for the APUs that are starved for memory bandwidth that sounds like it could give a great performance uplift.
Lucas
2024-08-08 07:08:39 +0000 UTC
