Kuantum (gölge mod)Quantum (shadow mode)
Gölge mod. Kuantum bileşenleri hiçbir uyarıyı ya da eşiği belirlemez. Simülatör (ideal ya da gürültülü) sonucu kuantum üstünlüğü değildir. Üstünlük iddiası yalnızca önceden kaydedilmiş, tutulan bir alt kümede, gerçek donanımda, eşleşik bootstrap güven aralığı sıfırı kuantum lehine dışlarsa (+0,02 AUC payıyla) 'aday' olarak yazılır ve insan incelemesi gerekir.Shadow mode. Quantum components never set an alert or threshold. A simulator result (ideal or noisy) is not evidence of a hardware benefit. An advantage claim is only written as a 'candidate' when a pre-registered held-out subset, run on real hardware, has a paired bootstrap CI that excludes zero in quantum's favour (with a +0.02 AUC margin), and it then needs human review.
Son gölge koşusuLatest shadow run · 2026-11-02
QF-MY-1: 6 kübit, 2 katman VQC; tutulan dönemden önceden kayıtlı dengeli alt küme (16 sel + 16 sel-yok gün). QF-MY-2: 8 eyalet / K=3 sensör yerleşimi, QAOA p=1 (28.10.2026'dan beri XY + Dicke); kesin MILP ile kıyas.QF-MY-1: 6-qubit, 2-layer VQC; pre-registered balanced held-out subset (16 flood + 16 no-flood days). QF-MY-2: 8-state / K=3 sensor siting, QAOA p=1 (XY + Dicke since 28 Oct 2026), compared with exact MILP.
| tier | backend | kind | fallback | status | vqc_auc_subset | twin_auc_subset | diff_ci | qaoa_gap_pct | qaoa_p_opt | shots | decision |
|---|
| ideal_sim | aer_simulator | local_simulator | none | done | 0.7773 | 0.832 | [-0.1253, 0.0] | 0.0 | 0.0485 | 18384 | golge — benzetim, ustunluk sayilmaz |
| noisy_sim | aer_noisy[ibm_marrakesh] | local_simulator | none | done | 0.8164 | 0.832 | [-0.085, 0.0437] | 0.0 | 0.026 | 18384 | golge — benzetim, ustunluk sayilmaz |
| hardware | aer_simulator (yerel) | simulator | IBM kalan kota 54 s < 60 s | fallback_local_sim | 0.6 | 0.76 | ['', ''] | 0.0 | 0.04 | | golge — tek gunluk kucuk kosu, ustunluk degerlendirmesi yapilmaz |
Tüm tutulan dönem (benzetim)Full held-out period (simulation): AUC VQC 0.8196 · en iyi ikizbest twin (logistic) 0.8475 · fark %95 GAdiff 95% CI [-0.047, -0.0088] → gölgeshadow
Ön-kayıtlı donanım çalışması (Ek 2 / Ek 3)Pre-registered hardware study (Amendment 2 / 3) · 2026-10-28
Stilize havza hücreleri doğrudan çip kubitlerine yerleştirildi (100–148 kubit, ibm_kingston; 32 kubit, Origin WK_C180). 60 önceden hash'lenmiş örnek × {QAOA p=1, p=2, |+⟩ kontrol} × 4096 atış. Ölçüt: ortalama enerji / optimum (rastgele = 0).Stylised basin cells were mapped directly onto chip qubits (100–148 qubits on ibm_kingston; 32 qubits on Origin WK_C180). 60 pre-hashed instances × {QAOA p=1, p=2, |+⟩ control} × 4096 shots. Metric: mean energy / optimum (random = 0).
| provider | backend | job_id | circuits | shots | submitted | qpu_s |
|---|
| IBM | ibm_kingston | dbgrr9pt9uhs73c8rsu0 | 60 | 4096 | 2026-10-28 12:17:28 TRT | 67 |
| IBM | ibm_kingston | dbgrs0hq260s73btuq90 | 60 | 4096 | 2026-10-28 12:18:58 TRT | 67 |
| IBM | ibm_kingston | dbgrsmpt9uhs73c8rv40 | 60 | 4096 | 2026-10-28 12:20:27 TRT | 67 |
| Origin | WK_C180 | 02AE40918F5BA68DBEABF1ED3D3DDA03 | 7 | 4096 | 2026-10-28 | 8.7 (qpuRunTime) |
| cls | N | instances | p1_AR | p1_minus_control_CI | p2_AR | p2_minus_p1_CI | hw_over_predicted | ideal_p1 | best_of_4096_p2 | random_best_of_4096 |
|---|
| HX2 | 100 | 10 | 0.42430670227233513 | [0.4200766354912568, 0.42878920996477726] | 0.5302564651536812 | [0.10127499062001498, 0.11135080060048279] | 0.9525311827289744 | 0.47163018631719955 | 0.8126739197468502 | 0.3644692354412557 |
| HX2 | 133 | 10 | 0.421085829287622 | [0.41541233119743914, 0.42657934151157983] | 0.5211602908513626 | [0.09377264285318777, 0.10638417870018566] | 0.9440182323088793 | 0.47699727139001435 | 0.798638095537702 | 0.3316156123986419 |
| HX2 | 148 | 10 | 0.42185219665268325 | [0.4182730755484316, 0.42501707009468054] | 0.5222764515666722 | [0.09641035838708309, 0.10517436063908016] | 0.9499040959213276 | 0.4775205280507639 | 0.7770812264502845 | 0.32659925214548974 |
| HX3 | 100 | 10 | 0.34910838520341475 | [0.3390871121550845, 0.35950273809902733] | 0.4582467765712496 | [0.10544282750405211, 0.11320311933062051] | 0.9522121006287122 | 0.4045110737196659 | 0.76428315503516 | 0.38881395221597875 |
| HX3 | 133 | 10 | 0.3455074944877998 | [0.3405130398244604, 0.3533596234060891] | 0.45198632900182656 | [0.1036459711271744, 0.10955729875035546] | 0.9488313382933672 | 0.4079560990858016 | 0.7113940452230033 | 0.3388498314084147 |
| HX3 | 148 | 10 | 0.3375872577608265 | [0.3307126046062877, 0.3447259983955921] | 0.44161166432975596 | [0.10107754615031771, 0.10720648244470056] | 0.952361054919985 | 0.3987756076304136 | 0.696200508313834 | 0.3160388667615843 |
| instance | hw_AR | control | diff_CI | predicted | ratio | mitigated_AR |
|---|
| WK3_HX2_N32_s4000 | 0.2430527098300403 | -0.0002009722463547138 | [0.23492730799636, 0.25165891618932357] | 0.2282679558623191 | 1.0647692923514795 | 0.29932342932546063 |
| WK3_HX3_N32_s4000 | 0.1805898416545767 | -0.0004306452036903837 | [0.17281866911666097, 0.1891771958512143] | 0.14884852915357452 | 1.2132457249090656 | 0.22624651430609796 |
| WK3_HX2_N32_s4001 | 0.2612101050932468 | 0.016770560555047522 | [0.23642030031427952, 0.2526581733169363] | 0.2303745635622118 | 1.1338495928293228 | 0.31409634217477655 |
| WK3_HX3_N32_s4001 | 0.19142425006349628 | 0.014756933945719745 | [0.16876111805103064, 0.18445806751743501] | 0.1615071844255303 | 1.1852367481011998 | 0.23484552325148636 |
Q-Flood'un stilize havza yerleşim problemleri 100–148 kubitlik gerçek kuantum işlemcide (IBM Heron) rastgele örneklemeden istatistiksel olarak anlamlı biçimde daha iyi çözümler üretti ve sonuçlar önceden kaydedilmiş gürültü tahminiyle uyumludur. Origin Wukong'da (32 kubit, okuma hatası modeli önceden kaydedilmiş) da aynı sonuç alındı. Aynı problemleri klasik kesin çözücüler milisaniyede optimal çözer; bu, klasik yöntemlere üstünlük kanıtı değildir. Donanım hiçbir örnekte optimuma ulaşmadı.Q-Flood's stylised basin-siting problems produced solutions statistically significantly better than random sampling on a 100–148-qubit real quantum processor (IBM Heron), consistent with the pre-registered noise prediction. The same holds on Origin Wukong (32 qubits, with a pre-registered readout-error model). Exact classical solvers solve the same problems optimally in milliseconds; this is not evidence of superiority over classical methods. The hardware never reached the optimum on any instance.
QAOA düzeltmesiQAOA fix · 2026-10-28
Dürüst not: 28 Ekim 2026'ya kadar kullanılan QAOA devresi rastgele seçimden kötüydü (ideal benzetimde geçerli çözüm %12,2; rastgele %21,9). Neden bir parametre arama hatasıydı: p=1 ızgarası β'yı yalnızca [0,05, π/2] aralığında tarıyordu, iyi bölge ise β<0'daydı. Ayrıca ceza terimli form, geçerli çözümler içinde amacı değil yalnızca kısıtı öğreniyordu. Yeni devre K=3 ağırlıklı Dicke durumundan başlar ve XY-halka karıştırıcı kullanır; bu karıştırıcı seçilen sensör sayısını korur. Maliyet yalnızca amaçtır (ceza yok). Parametreler γ∈[−π,π], β∈[−π/2,π/2] tam ızgara + COBYLA ile seçilir. Günlük donanım sınırları değişmedi (1 iş/gün, toplam 1000 shot, 60 s, yalnız open-instance). Bu bir doğruluk düzeltmesidir; klasik yöntemlere üstünlük kanıtı değildir; klasik MILP aynı problemi milisaniyede kesin çözer.Honest note: until 28 Oct 2026 the QAOA circuit was worse than random selection (ideal simulation: valid solutions 12.2% vs 21.9% random). The cause was a parameter-search bug: the p=1 grid only scanned β in [0.05, π/2], while the good region is β<0. In addition, the penalty form learned only the constraint, not the objective, among valid solutions. The new circuit starts from a weight-3 Dicke state and uses an XY ring mixer that preserves the number of selected sensors. The cost is the objective only (no penalty). Parameters come from a full grid γ∈[−π,π], β∈[−π/2,π/2] + COBYLA. Daily hardware limits are unchanged (1 job/day, 1000 shots total, 60 s, open-instance only). This is a correctness fix, not evidence of superiority over classical methods: classical MILP solves the same problem exactly in milliseconds.
| tier | circuit | feasible | p_opt | random_feasible | random_p_opt_feasible |
|---|
| ideal | eski / old: X + ceza / penalty | 0.1223 | 0.0022 | 0.2188 | 0.0179 |
| ideal | yeni / new: XY + Dicke(K=3) | 1.0 | 0.0494 | 0.2188 | 0.0179 |
| noisy · fake_fez | eski / old: X + ceza / penalty | 0.1515 | 0.004 | 0.2188 | 0.0179 |
| noisy · fake_fez | yeni / new: XY + Dicke(K=3) | 0.605 | 0.026 | 0.2188 | 0.0179 |
| noisy · fake_marrakesh | eski / old: X + ceza / penalty | 0.1512 | 0.0018 | 0.2188 | 0.0179 |
| noisy · fake_marrakesh | yeni / new: XY + Dicke(K=3) | 0.6943 | 0.0338 | 0.2188 | 0.0179 |
| backend | layout | old_depth | old_2q | new_depth | new_2q |
|---|
| fake_fez | [126, 125, 124, 123, 136, 143, 142, 141] | 322 | 152 | 467 | 176 |
| fake_marrakesh | [7, 8, 9, 10, 11, 12, 13, 14] | 322 | 152 | 464 | 176 |
| ibm_kingston | [91, 98, 111, 110, 109, 118, 129, 128] | 322 | 152 | 463 | 176 |
Transpile: optimization_level 3, ibm_kingston hedefi salt-okunur alındı (iş gönderilmedi). Gürültülü satırlar FakeFez/FakeMarrakesh gürültü modelleriyle 4000 shot.Transpile: optimization_level 3; the ibm_kingston target was read-only (no job submitted). Noisy rows: FakeFez/FakeMarrakesh noise models, 4000 shots.
Ön-kayıtlı avantaj çalışması — pilotPre-registered advantage study — pilot · 2026-10-28
Ön-kayıt kilitlendi (gövde sha256 Pre-registration locked (body sha256 9be239ebff2cd64c…). Pilot (2 boyut × 2 örnek × 4 devre × 4096 shot, N=100/133) ibm_kingston için hazırlandı ama gönderilmedi: derin devreler nedeniyle tahmini kotaThe pilot (2 sizes × 2 instances × 4 circuits × 4096 shots, N=100/133) was prepared for ibm_kingston but not submitted: because the circuits are deep, the estimated quota use is 67.5 s, onaylı üst sınır 30 s. Harcanan QPU süresi: 0 s. Kalibrasyona göre tahmini devre doğruluğu ≈ 0 (N=100'de ≈ e^−105), yani donanım çıktısı gürültüden ayırt edilemezdi. İzin verilen ifade yalnızca: “Q-Flood kuantuma hazır (quantum-ready)”; klasik yöntemlere üstünlük iddia edilmez.against an approved cap of 30 s. QPU time spent: 0 s. Calibration-based circuit fidelity ≈ 0 (≈ e^−105 at N=100), so hardware output would be indistinguishable from noise. Allowed wording is only: “Q-Flood is quantum-ready”; no superiority over classical methods is claimed.
İlk gerçek donanım koşusuFirst real-hardware run · 2026-10-27
Kullanıcı onaylı tek iş: 32 VQC + 1 QAOA devresi tek SamplerV2 işinde; devre başına 1000 shot; max_execution_time 60 s. Fiziksel kübitlerUser-approved single job: 32 VQC + 1 QAOA circuits in one SamplerV2 job; 1000 shots per circuit; max_execution_time 60 s. Physical qubits: VQC [148, 147, 151, 152, 150, 149] · QAOA [118, 129, 108, 109, 117, 107, 105, 106] · derinlikdepth 95 / 322
| provider | backend | instance | job_id | status | circuits | shots | created | running | finished | queue_s | qpu_s | remaining_s | cost |
|---|
| IBM | ibm_kingston | open-instance | dbgf8kgiudkc73e4cvr0 | DONE | 33 | 1000/circ · 33000 | 2026-10-27 21:58:26 TRT | 2026-10-27 21:58:28 TRT | 2026-10-27 21:59:01 TRT | 1.9 | 11 | 563 | 0 USD (Open Plan ucretsiz kota; 600 s / 28 gun) |
| Origin | WK_C180 | | — | sunucu reddetti / rejected by server: QPU time is insufficient. Please purchase more. | | | 2026-10-27 22:14:39 | | | | | | |
| tier | vqc_auc_subset | twin_auc_subset | diff_ci | agree_twin | vqc_tvd_vs_exact | qaoa_feasible | qaoa_p_opt | qaoa_gap_pct | qaoa_tvd_vs_exact |
|---|
| ideal_aer | 0.8047 | 0.832 | [-0.0994, 0.0329] | 0.9062 | 0.0699 | 0.124 | 0.001 | 0.0 | 0.1956 |
| noisy_kingston_model | 0.7734 | 0.832 | [-0.1389, 0.0039] | 0.9062 | 0.1038 | 0.128 | 0.003 | 0.0 | 0.1995 |
| hardware_ibm_kingston | 0.7812 | 0.832 | [-0.128, 0.0182] | 0.8125 | 0.1535 | 0.074 | 0.001 | 0.0 | 0.3063 |
| pair | vqc_tvd_mean | vqc_tvd_max | qaoa_tvd |
|---|
| ideal_aer ↔ noisy_kingston_model | 0.1228 | 0.16 | 0.276 |
| ideal_aer ↔ hardware_ibm_kingston | 0.1711 | 0.222 | 0.358 |
| noisy_kingston_model ↔ hardware_ibm_kingston | 0.1402 | 0.18 | 0.357 |
Tek koşu, 32 örneklik alt küme: güven aralığı sıfırı içeriyor → kuantum üstünlüğü yok, gölge modda kalır. Gürültülü katman: AerSimulator.from_backend(ibm_kingston), aynı fiziksel kübitler. QAOA p=1 uygun-çözüm oranı rastgele tabanın (C(8,3)/256 = 0,219) altında (eski devre; parametre arama hatası, 28.10.2026'da düzeltildi). Origin WK_C180: tek toplu iş gönderildi, sunucu yetersiz QPU süresi nedeniyle reddetti; iş oluşmadı.Single run, 32-sample subset: the CI includes zero → no evidence of a benefit over classical models; stays in shadow mode. Noisy tier: AerSimulator.from_backend(ibm_kingston) on the same physical qubits. QAOA p=1 feasible rate is below the random baseline (C(8,3)/256 = 0.219) (old circuit; parameter-search bug, fixed on 28 Oct 2026). Origin WK_C180: the single batch was submitted but the server rejected it for insufficient QPU time; no job was created.
Üçlü karşılaştırma: IBM kingston · Wukong · benzetimlerThree-way: IBM kingston · Wukong · simulations
| tier | auc | ci_vs_twin | agree_twin | agree_exact | vqc_tvd_exact | qaoa_valid | qaoa_p_best | qaoa_tvd_exact |
|---|
| ideal Aer | 0.8047 | [-0.0994, 0.0329] | 0.9062 | 1.0 | 0.0699 | 0.124 | 0.001 | 0.1956 |
| noisy · kingston model | 0.7734 | [-0.1389, 0.0039] | 0.9062 | 1.0 | 0.1038 | 0.128 | 0.003 | 0.1995 |
| noisy · Wukong local approx. | 0.832 | [-0.0802, 0.0883] | 0.5 | 0.5938 | 0.4075 | 0.201 | 0.008 | 0.2688 |
| HW · IBM kingston | 0.7812 | [-0.128, 0.0182] | 0.8125 | 0.9062 | 0.1535 | 0.074 | 0.001 | 0.3063 |
| HW · Origin WK_C180 | — | sunucu reddetti / rejected: QPU time is insufficient. Please purchase more. | | | | | | |
Rastgele taban: geçerli çözüm payı 0,219 (C(8,3)/256); en iyi çözüm olasılığı 1/256 = 0,0039. Wukong satırı gerçek donanım değil, WK_C180 kalibrasyon medyanlarından yerel gürültü yaklaşımıdırRandom baseline: valid-solution share 0.219 (C(8,3)/256); best-solution probability 1/256 = 0.0039. The Wukong row is not hardware but a local noise approximation from WK_C180 calibration medians: Aer depolarizing u3 λ=0.0048 (≈2 RPhi), cz λ=0.0276, simetrik okuma hatasi 0.0714; dogrusal zincir yonlendirme (u3+cz) — YEREL YAKLASIM, Origin bulut benzetimi degil
İş kökeni (provenance)Job provenance
| provider | mode | backend | kind | job | job_id | circuits | qubits | shots | submitted | queue_s | status |
|---|
| local_aer | simulator | aer_simulator | local_simulator | qf_my_1_vqc_subset | local-545345f0 | 32 | 6 | 512 | 2026-11-02T06:33:17+03:00 | 0.0 | done |
| local_aer | simulator | aer_simulator | local_simulator | qf_my_2_qaoa_xy_dicke_p1 | local-d465769c | 1 | 8 | 2000 | 2026-11-02T06:33:19+03:00 | 0.0 | done |
| ibm | noisy_simulator | aer_noisy[ibm_marrakesh] | local_simulator | qf_my_1_vqc_subset | local-06eb11ce | 32 | 6 | 512 | 2026-11-02T06:33:34+03:00 | 0.0 | done |
| ibm | noisy_simulator | aer_noisy[ibm_marrakesh] | local_simulator | qf_my_2_qaoa_xy_dicke_p1 | local-af58aa46 | 1 | 8 | 2000 | 2026-11-02T06:33:43+03:00 | 0.0 | done |
Gölge geçmişiShadow history
| date | tier | provider | backend | kind | auc_subset | auc_twin_subset | ci_lo | ci_hi | qaoa_gap_pct | qaoa_p_opt | status |
|---|
| 2026-11-02 | hardware | ibm | aer_simulator (yerel) | simulator | 0.6 | 0.76 | | | 0.0 | 0.04 | fallback_local_sim |
| 2026-11-02 | noisy_sim | ibm | aer_noisy[ibm_marrakesh] | local_simulator | 0.8164 | 0.832 | -0.085 | 0.0437 | 0.0 | 0.026 | done |
| 2026-11-02 | ideal_sim | local_aer | aer_simulator | local_simulator | 0.7773 | 0.832 | -0.1253 | 0.0 | 0.0 | 0.0485 | done |
| 2026-11-01 | hardware | ibm | aer_simulator (yerel) | simulator | 0.6 | 0.76 | | | 0.0 | 0.04 | fallback_local_sim |
| 2026-11-01 | noisy_sim | ibm | aer_noisy[ibm_marrakesh] | local_simulator | 0.8125 | 0.832 | -0.0863 | 0.0336 | 0.0 | 0.0265 | done |
| 2026-11-01 | ideal_sim | local_aer | aer_simulator | local_simulator | 0.7773 | 0.832 | -0.1253 | 0.0 | 0.0 | 0.0485 | done |
| 2026-10-31 | hardware | ibm | ibm_kingston | hardware | 0.64 | 0.76 | | | 0.0 | 0.015 | done |
| 2026-10-31 | noisy_sim | ibm | aer_noisy[ibm_fez] | local_simulator | 0.8125 | 0.832 | -0.0908 | 0.0516 | 0.0 | 0.022 | done |
| 2026-10-31 | ideal_sim | local_aer | aer_simulator | local_simulator | 0.7773 | 0.832 | -0.1253 | 0.0 | 0.0 | 0.0485 | done |
| 2026-10-30 | hardware | ibm | ibm_kingston | None | 0.6 | 0.76 | | | 0.0 | 0.0175 | done_earlier_today |
| 2026-10-30 | noisy_sim | ibm | aer_noisy[ibm_fez] | local_simulator | 0.8086 | 0.832 | -0.0947 | 0.0324 | 0.0 | 0.032 | done |
| 2026-10-30 | ideal_sim | local_aer | aer_simulator | local_simulator | 0.7773 | 0.832 | -0.1253 | 0.0 | 0.0 | 0.0485 | done |
| 2026-10-29 | hardware | ibm | ibm_kingston | hardware | | | | | | | pending |
| 2026-10-29 | noisy_sim | ibm | aer_noisy[ibm_fez] | local_simulator | 0.8203 | 0.832 | -0.0928 | 0.0627 | 0.0 | 0.027 | done |
| 2026-10-29 | ideal_sim | local_aer | aer_simulator | local_simulator | 0.7773 | 0.832 | -0.1253 | 0.0 | 0.0 | 0.0485 | done |
| 2026-10-28 | noisy_sim | ibm | aer_noisy[ibm_fez] | local_simulator | 0.8086 | 0.832 | -0.0939 | 0.0336 | 0.0 | 0.0025 | done |
| 2026-10-28 | ideal_sim | local_aer | aer_simulator | local_simulator | 0.7773 | 0.832 | -0.1253 | 0.0 | 0.0 | 0.002 | done |
| 2026-10-27 | hardware | ibm | ibm_kingston | cloud_hardware | 0.7812 | 0.832 | -0.128 | 0.0182 | 0.0 | 0.001 | done |
| 2026-10-27 | noisy_sim | ibm | aer_noisy[ibm_kingston] | local_simulator | 0.7734 | 0.832 | -0.1389 | 0.0039 | 0.0 | 0.003 | done |
| 2026-10-27 | hardware | local_aer | aer_noisy[fake_fez] | local_simulator | | | | | | | not_submitted |
| 2026-10-27 | noisy_sim | local_aer | aer_noisy[fake_fez] | local_simulator | 0.7891 | 0.832 | -0.1212 | 0.0227 | 0.0 | 0.002 | done |
| 2026-10-27 | ideal_sim | local_aer | aer_simulator | local_simulator | 0.7773 | 0.832 | -0.1253 | 0.0 | 0.0 | 0.002 | done |
Görünen kuantum arka uçları (salt-okunur kontrol)Visible quantum backends (read-only check)
| provider | name | qubits | status | pending_jobs | median_2q_error | median_readout_error |
|---|
| IBM | ibm_fez | 156 | active | 6 | 0.002527 | 0.010742 |
| IBM | ibm_kingston | 156 | active | 5 | 0.001714 | 0.010986 |
| IBM | ibm_marrakesh | 156 | maintenance | 13 | 0.002529 | 0.026306 |
| provider | name | kind | qubits | usable_qubits | status | median_2q_fidelity |
|---|
| Origin | HanYuan_01 | QPU | | | offline | |
| Origin | PQPUMESH8 | QPU | 3 | 3 | available | 0.9593 |
| Origin | WK_C180 | QPU | 180 | 166 | available | 0.9793 |
| Origin | WK_C180_2 | QPU | 180 | 17 | available | 0.99095 |
| Origin | WK_C180_3 | QPU | 180 | 30 | available | 0.97855 |
| Origin | full_amplitude | simulator | 35 | | available | |
| Origin | partial_amplitude | simulator | 68 | | available | |
| Origin | single_amplitude | simulator | 200 | | available | |
IBM bulut simülatörleri 15 Mayıs 2024'te emekli edildi; 'ideal' ve 'gürültülü' katmanlar yerel Qiskit Aer ile (gürültülü katman seçilen cihazın canlı kalibrasyonundan) koşar. Origin Quantum'da tam genlik bulut simülatörü kullanılabilir.IBM cloud simulators were retired on 15 May 2024; the 'ideal' and 'noisy' tiers run on local Qiskit Aer (the noisy tier uses the selected device's live calibration). Origin Quantum offers a full-amplitude cloud simulator.